Goodheart KA

Low Firing Temperature Absorption Chiller System
by

KEVIN A. GOODHEART

A thesis submitted in partial fulfillment
of the requirements for the degree of

MASTER OF SCIENCE

(MECHANICAL ENGINEERING)

at the

UNIVERSITY OF WISCONSIN MADISON

DECEMBER 2000

Copyright 2000 Kevin A. Goodheart

Approved by

_______________________________

Professor Sanford A. Klein
December 12, 2000

ABSTRACT

Two types of the absorption chillers, the single and half effect cycles, can operate using low
temperature hot water. The advantage of the single effect over the half effect cycle is the
higher COP but the single effect has a narrower temperature operating range. The half effect
has the capability of using lower hot water temperature but the lower COP increases
operating cost and requires investing a larger cooling tower. A detailed computer model was
written for the single and half effect cycles based on heat transfer coefficients for the inside
and outside tubes of each component [generator, absorber, condenser, and evaporator],
energy, mass, salt balances, and rate equations. The single effect component model was
calibrated with known data from a US absorption chiller manufacturer. The cooling tower
was modeled using the analogy approach, calibrated and validated with performance data
from a cooling tower manufacturer. Capacity and dollars per ton were used to determine
lower limits on the firing temperature. The results show that at 225
o
F (107 C) at 2000 gpm
(7570 L/min) of hot water, the cost to maintain capacity starts to change for the single effect
cycle. At 205
o
F (99 C) the cost to maintain capacity increases rapidly. The half effect cycle
can maintain capacity at temperatures as low as 185
o
F (82 C) hot water at 2000 gpm (7570
L/min) without a large increase in capital investment. The capital cost for the half effect
chiller system is 200 $/ton (57 $/kW) more than the single effect, using hot water
temperatures above 200
o
F (93 C). The single effect cycle can only be competitive with an
electric centrifugal chiller if the heat source is free or a combination of high electrical cost
with a low cost of heat.

v

The half effect cycle can be competitive with the single effect if the waste heat is free and the
temperature is below 200
o
F (93 C) or has a low flow rate in the range of 1000 gpm (3785
L/min).
vi

ACKNOWLEDGMENTS

The first acknowledgment is to recognize Ken Schultz for making this project possible and
being available to answer questions and supply information on absorption chillers. I would
like to thank the Trane Company for their financial support for the project [The facts and
findings expressed in this thesis do not necessarily reflect the views held by the Trane
Company].

My second acknowledgment is to recognize the importance of using EES in modeling
systems. With this model I can solve for the capacity given the UA of the generator and in a
second I can solve for the UA of the generator given the capacity. The above procedure is
done by changing the arrangement of how 700 equations are solved. I came to this lab with a
set of books and a HP calculator and I am leaving this lab with a set of books and the ability
to model systems using EES. It has been a great pleasure to work with Sanford Klein the
developer of EES on this project and as his student in Thermodynamics. Without EES I
would not have had the inspiration to examine the different facets of an absorption chiller
and an absorption chiller system.

I would also like to acknowledge William Beckman the director of the Solar Energy lab for
making it possible to have this place to study and learn. The way in which he asks questions
is encouraging because it promotes one to think about what they are doing.

vii

I would like to thank my colleagues in the lab for letting me use the same computer, listening
to my problems, hearing my cheers when something worked, and all around making my time
in the SEL lab a memorable occasion.

And finally

Gratiam soli agere volo
quia facit dies tempus degendos
quod litteras in annis
meas faciet obsoletas futuris
Latin translation by Rdiger Spahl
viii

TABLE OF CONTENTS
ABSTRACT............................................................................................................................ V
ACKNOWLEDGMENTS.................................................................................................. VII
TABLE OF CONTENTS ..................................................................................................... IX
LIST OF FIGURES...........................................................................................................XIII
LIST OF TABLES........................................................................................................... XVII
NOMENCLATURE........................................................................................................... XIX
CHAPTER 1............................................................................................................................ 1
INTRODUCTION................................................................................................................... 1
1.1 WHY ABSORPTION? ................................................................................................... 2
1.2 WHAT IS ABSORPTION?.............................................................................................. 2
1.3 HISTORY OF LARGE TONNAGE ABSORPTION UNITS ..................................................... 5
1.4 WHERE ARE COMMERCIAL ABSORPTION UNITS FOUND?............................................. 6
1.5 WHO MAKES ABSORPTION UNITS? ............................................................................. 7
CHAPTER 2............................................................................................................................ 9
REVIEW OF PUBLISHED WORK ON LOW TEMPERATURE HEAT SOURCE
APPLICATIONS .................................................................................................................... 9
2.1 OVERVIEW................................................................................................................. 9
2.2 LOW TEMPERATURE HEAT SOURCE APPLICATIONS USING THE SINGLE-EFFECT
ABSORPTION UNIT ................................................................................................................ 10
2.3 HALF-EFFECT ABSORPTION UNIT.............................................................................. 16
2.4 SUMMARY................................................................................................................ 21
CHAPTER 3.......................................................................................................................... 23
SINGLE-EFFECT ABSORPTION MODEL DEVELOPMENT .................................... 23
3.1 OVERVIEW............................................................................................................... 23
3.2 MODEL INPUT PARAMETERS.................................................................................... 25
3.2.1 Generator............................................................................................................ 25
3.2.2 Condenser ........................................................................................................... 25
3.2.3 Evaporator .......................................................................................................... 26
3.2.4 Absorber.............................................................................................................. 26
3.2.5 Low-Temperature Heat Exchanger..................................................................... 26
3.3 ASSUMPTIONS.......................................................................................................... 26
ix

3.3.1 Overall ................................................................................................................ 26
3.3.2 Generator............................................................................................................ 27
3.3.3 Condenser ........................................................................................................... 27
3.3.4 Evaporator .......................................................................................................... 27
3.3.5 Absorber.............................................................................................................. 27
3.4 MODEL DEVELOPMENT FOR STEAM FIRED ABSORPTION UNIT................................... 27
3.4.1 Overall ................................................................................................................ 27
3.4.2 Generator............................................................................................................ 28
3.4.3 Absorber.............................................................................................................. 31
3.4.4 Condenser ........................................................................................................... 35
3.4.5 Evaporator .......................................................................................................... 37
3.5 CONTROL STRATEGY ............................................................................................... 39
3.6 CALIBRATION........................................................................................................... 42
3.7 MODEL DEVELOPMENT: HOT WATER FIRED GENERATOR....................................... 51
3.7.1 Difference between steam fired and hot water fired generator .......................... 51
3.7.2 Inside Heat Transfer Coefficient......................................................................... 52
3.7.3 Outside Heat Transfer Coefficient ...................................................................... 52
CHAPTER 4.......................................................................................................................... 59
MODEL DEVELOPMENT OF HALF EFFECT CYCLE............................................... 59
4.1 OVERVIEW............................................................................................................... 59
4.2 INPUTS ..................................................................................................................... 63
4.2.1 HTGenerator....................................................................................................... 63
4.2.2 LTGenerator ....................................................................................................... 63
4.2.3 Condenser ........................................................................................................... 63
4.2.4 Evaporator .......................................................................................................... 64
4.2.5 HC Absorber ....................................................................................................... 64
4.2.6 LC Absorber........................................................................................................ 64
4.2.7 High-Temperature Heat Exchanger ................................................................... 64
4.2.8 Low-Temperature Heat Exchanger..................................................................... 65
4.3 ASSUMPTIONS.......................................................................................................... 65
CHAPTER 5.......................................................................................................................... 67
COOLING TOWER............................................................................................................. 67
5.1 OVERVIEW............................................................................................................... 67
5.2 ANALOGY APPROACH.............................................................................................. 69
5.3 FAN POWER ............................................................................................................. 71
5.4 OVERALL TOWER HEAT TRANSFER COEFFICIENT.................................................... 72
5.5 CALIBRATION/PREDICTION...................................................................................... 74
CHAPTER 6.......................................................................................................................... 77
ECONOMICS ....................................................................................................................... 77
6.1 OVERVIEW............................................................................................................... 77
x

6.2 MATERIAL / INSTALLATION COST ASSUMPTIONS .................................................... 78
6.2.1 Absorption chiller ............................................................................................... 78
6.2.2 Electric Centrifugal Chiller ................................................................................ 79
6.2.3 Cooling Tower .................................................................................................... 80
6.3 OPERATING COST ASSUMPTIONS ............................................................................. 80
6.4 LIFE CYCLE COST APPROACH.................................................................................. 81
CHAPTER 7.......................................................................................................................... 85
SINGLE-EFFECT ABSORPTION CHILLER SYSTEM OPERATION....................... 85
7.1 OVERVIEW............................................................................................................... 85
7.2 OPTIMIZATION OF THE ABSORPTION CHILLER SYSTEM FOR A GIVEN HEAT SOURCE
TEMPERATURE ..................................................................................................................... 86
7.3 EFFECT OF HEAT SOURCE TEMPERATURES............................................................... 95
7.4 SENSITIVITY ANALYSIS ON UA OF EACH COMPONENT............................................. 97
7.5 SENSITIVITY ANALYSIS ON TUBE COST AND GENERATOR OUTSIDE HEAT TRANSFER
COEFFICIENT. ..................................................................................................................... 106
7.6 LIFE CYCLE SAVINGS COMPARISON BETWEEN SINGLE-EFFECT ABSORPTION AND
VAPOR COMPRESSION CHILLERS FOR DIFFERENT HEAT SOURCE TEMPERATURES................ 108
7.7 SUMMARY.............................................................................................................. 114
CHAPTER 8........................................................................................................................ 117
HALF-EFFECT ABSORPTION CHILLER OPERATION.......................................... 117
8.1 OVERVIEW............................................................................................................. 117
8.2 DESIGNING THE HALF-EFFECT CYCLE................................................................... 117
8.3 EFFECT OF HEAT SOURCE TEMPERATURE.............................................................. 126
8.4 DOLLARS PER TON EFFECT DUE TO SPLIT OR CONTINUOUS HEAT SOURCE FLOW RATE
INTO THE GENERATORS ...................................................................................................... 129
8.5 LIFE CYCLE SAVINGS COMPARISON BETWEEN HALF-EFFECT ABSORPTION AND VAPOR
COMPRESSION CHILLERS .................................................................................................... 130
8.6 VARIATION OF CAPACITY DUE TO DIFFERENT OPERATING CONDITIONS AND DESIGNS.
133
8.7 SUMMARY.............................................................................................................. 139
CHAPTER 9........................................................................................................................ 141
COMPARISON BETWEEN HALF-EFFECT AND SINGLE-EFFECT ABSORPTION
CHILLERS.......................................................................................................................... 141
9.1 OVERVIEW............................................................................................................. 141
9.2 DOLLAR PER TON FOR DIFFERENT HEAT SOURCE TEMPERATURES. ......................... 141
9.3 BREAK-EVEN COST COMPARISON. .......................................................................... 146
9.4 SUMMARY.............................................................................................................. 151
CHAPTER 10...................................................................................................................... 153
CONCLUSIONS AND RECOMMENDATIONS............................................................ 153
xi

10.1 CONCLUSIONS........................................................................................................ 153
10.2 RECOMMENDATIONS.............................................................................................. 155
References............................................................................................................................................................................. 157
Appendix A CD-ROM Files ............................................................................................................................................... 159

xii

LIST OF FIGURES
FIGURE 1-1. SCHEMATIC OF A CENTRIFUGAL CHILLER. ............................................................. 2
FIGURE 1-2. SCHEMATIC OF THE SINGLE-EFFECT ABSORPTION CHILLER. .................................. 3
FIGURE 1-3. SINGLE-EFFECT ABSORPTION CHILLER (US CHILLER MANUFACTURER, 2000). ... 4
FIGURE 2-1. SIMPLE CHP USING A GAS TURBINE (MELOCHE ET AL., 1996)............................ 13
FIGURE 2-2. FLOW CHART OF THE SE/HE CHILLER (SCHWEIGLER ET AL., 1999). .................. 18
FIGURE 3-1. SCHEMATIC OF THE SINGLE EFFECT ABSORPTION UNIT. .................................... 24
FIGURE 3-2. EMPIRICAL RELATIONSHIP BETWEEN OUTSIDE NUSSELT NUMBER AND FILM
REYNOLDS NUMBER FOR THE GENERATOR. ...................................................................... 30
FIGURE 3-3. NUSSELT NUMBER AS A FUNCTION FILM REYNOLDS NUMBER. ............................ 33
FIGURE 3-4. SUB-COOLING AS A FUNCTION OF EVAPORATOR AND ABSORBER INLET COOLING
TEMPERATURE. ................................................................................................................ 34
FIGURE 3-5. ABSORBER OUTLET SOLUTION FLOW RATE AS A FUNCTION OF CAPACITY AND
TEMPERATURE. ................................................................................................................ 40
FIGURE 3-6. COMPARISON BETWEEN MODEL AND DATA FOR CAPACITY ................................. 43
FIGURE 3-7. COMPARISON BETWEEN MODEL AND DATA FOR COP. ........................................ 44
FIGURE 3-8. COMPARISON BETWEEN MODEL AND DATA FOR UA GENERATOR. ....................... 45
FIGURE 3-9. COMPARISON BETWEEN MODEL AND DATA FOR UA ABSORBER........................... 45
FIGURE 3-10. COMPARISON BETWEEN MODEL AND DATA FOR UA CONDENSER. ..................... 46
FIGURE 3-11. COMPARISON BETWEEN MODEL AND DATA FOR UA EVAPORATOR. ................... 46
FIGURE 3-12. COMPARISON BETWEEN MODEL AND DATA FOR MASS FLOW RATE LEAVING THE
ABSORBER. ....................................................................................................................... 47
FIGURE 3-13. COMPARISON BETWEEN MODEL AND DATA FOR RE-CIRCULATION FLOW RATE... 47
FIGURE 3-14. COMPARISON BETWEEN MODEL AND DATA FOR REFRIGERANT FLOW RATE....... 48
FIGURE 3-15. COMPARISON BETWEEN MODEL AND DATA FOR STEAM FLOW RATE.................. 48
FIGURE 3-16. COMPARISON BETWEEN MODEL AND DATA FOR CONCENTRATION LEAVING
ABSORBER. ....................................................................................................................... 48
FIGURE 3-17. COMPARISON BETWEEN MODEL AND DATA FOR CONCENTRATION LEAVING
GENERATOR. .................................................................................................................... 48
FIGURE 3-18. COMPARISON BETWEEN MODEL AND DATA FOR CONCENTRATION ENTERING
ABSORBER. ....................................................................................................................... 49
FIGURE 3-19. CAPACITY AS A FUNCTION OF LOAD FOR DIFFERENT INLET ABSORBER COOLING
TEMPERATURES................................................................................................................ 50
FIGURE 3-20. COP AS A FUNCTION OF LOAD FOR DIFFERENT INLET ABSORBER COOLING
TEMPERATURE. ................................................................................................................ 50
FIGURE 3-21. CHANGE IN OUTSIDE HEAT TRANSFER COEFFICIENT AS A FUNCTION OF
GENERATOR EQUILIBRIUM PRESSURE (WANG ET AL., 1999). ........................................... 53
FIGURE 3-22. CHANGE IN EXITING CONCENTRATION AS A FUNCTION OF SPRAYING DENSITY
(WANG ET AL., 1999)....................................................................................................... 53
FIGURE 3-23. CHANGE IN OUTSIDE HEAT TRANSFER COEFFICIENT AS A FUNCTION OF
GENERATOR SPRAYING DENSITY WANG ET AL., (1999).................................................... 54
xiii

FIGURE 3-24. OUTSIDE HEAT TRANSFER COEFFICIENT AS A FUNCTION OF REYNOLDS NUMBER.
......................................................................................................................................... 56
FIGURE 3-25. COMPARISON BETWEEN OUTSIDE HEAT TRANSFER COEFFICIENT FOR DIFFERENT
CORRELATIONS. ............................................................................................................... 58
FIGURE 4-1. SCHEMATIC OF THE HALF-EFFECT CYCLE. ........................................................... 60
FIGURE 4-2. DHRING PLOT OF THE HALF-EFFECT CYCLE....................................................... 61
FIGURE 4-3. DHRING PLOT COMPARISON BETWEEN THE HALF-EFFECT AND SINGLE-EFFECT
CYCLE. ............................................................................................................................. 62
FIGURE 5-1. SCHEMATIC DIAGRAM OF A COOLING TOWER..................................................... 68
FIGURE 5-2. UA
TOWER
VERSUS AIR FLOW RATE FOR 10 RANGE, T
WO
=85
O
F , T
WB
=78
O
F,
VA
W
=2000 [GPM] ............................................................................................................ 73
FIGURE 5-3. FAN POWER VERSUS WET BULB TEMPERATURE [4000 GPM, NC9202]................ 75
FIGURE 5-4. WATER OUTLET TEMPERATURE VERSUS WET BULB TEMPERATURE [4000 GPM,
NC9202].......................................................................................................................... 75
FIGURE 7-1. SINGLE-EFFECT ABSORPTION CHILLER SYSTEM................................................. 86
FIGURE 7-2. BASE CASE CAPACITY AND COP VERSUS ENTERING HOT WATER TEMPERATURE.87
FIGURE 7-3. THE EFFECT OF CAPITAL DOLLARS PER TON FOR A FACTOR CHANGE IN UA. ....... 88
FIGURE 7-4. CONTOUR PLOT OF PV FOR 200 F ENTERING HOT WATER, P
1
=1.0...................... 89
1
=0.8...................... 90
1
=1.2...................... 90
FIGURE 7-7. CONTOUR PLOT OF PV FOR 200 F ENTERING HOT WATER, 0.1 $/KWH ELECTRICAL
COST ................................................................................................................................ 91
FIGURE 7-8. CONTOUR PLOT OF PV FOR 200 F ENTERING HOT WATER, 0.2 $/KWH ELECTRICAL
COST ................................................................................................................................ 92
2
=5 OR 10 YEAR LIFE
CYCLE. ............................................................................................................................. 93
FIGURE 7-10. CONTOUR PLOT OF PV FOR 230
O
F ENTERING HOT WATER. ............................... 94
FIGURE 7-11. CAPITAL DOLLARS PER TON AS A FUNCTION OF ENTERING HOT WATER.............. 95
FIGURE 7-12. UA GENERATOR AS A FUNCTION OF ENTERING HOT WATER. .............................. 96
FIGURE 7-13. THE EFFECT OF CAPITAL DOLLARS PER TON FOR A FACTOR CHANGE IN UA. ..... 98
FIGURE 7-14. CAPACITY EFFECT DUE TO CHANGING LTHX EFFECTIVENESS. ....................... 100
FIGURE 7-15. CAPITAL DOLLARS PER TON DUE TO CHANGING LTHX EFFECTIVENESS ......... 100
FIGURE 7-16. COMPARISON BETWEEN FINNED AND SMOOTH TUBES. .................................... 101
FIGURE 7-17. COST COMPARISON BETWEEN FINNED AND SMOOTH TUBES............................ 102
FIGURE 7-18. CAPACITY EFFECT BETWEEN FINNED AND SMOOTH TUBE WITH AND WITHOUT
SUPERHEAT. ................................................................................................................... 103
FIGURE 7-19. CAPITAL DOLLARS EFFECT BETWEEN FINNED AND SMOOTH TUBE WITH AND
WITHOUT SUPERHEAT..................................................................................................... 103
FIGURE 7-20. CAPITAL DOLLARS PER TON COMPARISON BETWEEN SINGLE-EFFECT ABSORPTION
CHILLER AND SYSTEM. ................................................................................................... 105
FIGURE 7-21. UNCERTAINTY ANALYSIS ON THE COST OF TUBES. .......................................... 106
FIGURE 7-22. UNCERTAINTY ANALYSIS ON THE GENERATOR OUTSIDE HEAT TRANSFER
COEFFICIENT. ................................................................................................................. 107
FIGURE 7-23. BREAK-EVEN COST OF SUPPLY HEAT TO THE GENERATOR OR HEAT EQUIPMENT.
....................................................................................................................................... 109
xiv

FIGURE 7-24. CAPACITY AND COP AS A FUNCTION OF HEAT SOURCE TEMPERATURE........... 110
FIGURE 7-25. UNCERTAINTY ANALYSIS IN P
1
....................................................................... 111
FIGURE 7-26. UNCERTAINTY ANALYSIS IN P
2
....................................................................... 111
FIGURE 7-27. COST OF HEAT SOURCE UNCERTAINTY IN P
2
FOR FULL AND HALF A YEAR OF
OPERATION..................................................................................................................... 112
FIGURE 7-28. CAPACITY COMPARISON BETWEEN LOW TEMPERATURE HOT WATER DESIGN AND
BASE CASE DESIGN. ........................................................................................................ 114
FIGURE 8-1. HC ABSORBER TUBES EFFECT ON DOLLARS PER TON. ....................................... 120
FIGURE 8-2. LC ABSORBER TUBES EFFECT ON DOLLARS PER TON. ....................................... 120
FIGURE 8-3. CONTOUR PLOT OF CAPITAL DOLLAR PER TONS FOR SERIES FLOW..................... 122
FIGURE 8-4. CONTOUR PLOT OF CAPITAL DOLLAR PER TONS FOR PARALLEL FLOW. .............. 123
FIGURE 8-5. CONTOUR PLOT OF CAPITAL DOLLARS PER TON FOR SERIES FLOW [185
O
F HEAT
SOURCE]......................................................................................................................... 124
FIGURE 8-6. EFFECT ON CAPITAL COST DUE TO CHANGING THE SIZE OF EACH COMPONENT. . 125
FIGURE 8-7. HT GENERATOR UA AS A FUNCTION OF ENTERING HOT WATER TEMPERATURE.126
FIGURE 8-8. LT GENERATOR UA AS A FUNCTION OF ENTERING HOT WATER TEMPERATURE. 127
FIGURE 8-9. CAPITAL DOLLARS PER TON FOR VARYING ENTERING HOT WATER TEMPERATURE.
....................................................................................................................................... 127
FIGURE 8-10. CAPITAL DOLLARS COMPARISON BETWEEN HALF-EFFECT ABSORPTION CHILLER
AND SYSTEM .................................................................................................................. 128
FIGURE 8-11. EFFECT ON CAPITAL COST FOR DIFFERENT FLOW ARRANGEMENT 2000 [GPM]. 129
FIGURE 8-12. EFFECT ON CAPITAL COST FOR DIFFERENT FLOW ARRANGEMENT 1500 [GPM]. 130
FIGURE 8-13. COMPARISON BETWEEN HALF-EFFECT CYCLE AND ELECTRIC CENTRIFUGAL
CHILLER. ........................................................................................................................ 131
FIGURE 8-14. CHANGE IN CAPACITY AND COP AS A FUNCTION OF ENTERING HOT WATER
FIRING TEMPERATURE. ................................................................................................... 132
FIGURE 8-15. CAPACITY EFFECT DUE TO DIFFERENT HOT WATER FLOW RATES..................... 133
FIGURE 8-16. COMPARISON OF DIFFERENT COOLING WATER FLOW ARRANGEMENTS............ 134
FIGURE 8-17. EFFECT ON CAPACITY DUE TO VARYING HC ABSORBER SOLUTION FLOW RATE.
....................................................................................................................................... 135
FIGURE 8-18. EFFECT ON CAPACITY DUE TO VARYING HC ABSORBER SOLUTION FLOW RATE.
....................................................................................................................................... 136
FIGURE 8-19. SIMULTANEOUS REDUCTION IN BOTH INTERNAL SOLUTION FLOW RATE.......... 137
FIGURE 8-20. EFFECT ON CAPACITY DUE TO VARYING COOLING WATER TEMPERATURE. ...... 138
FIGURE 9-1. CAPITAL DOLLARS PER TON COST COMPARISON BETWEEN SINGLE AND HALF-
EFFECT CHILLER. ............................................................................................................ 142
FIGURE 9-2. SYSTEM DOLLARS PER TON COST COMPARISON BETWEEN SINGLE AND HALF-
EFFECT CHILLER. ............................................................................................................ 142
FIGURE 9-3. CAPITAL DOLLARS PER TON FOR DIFFERENT HOT WATER FLOW RATES.............. 144
FIGURE 9-4. SYSTEM DOLLARS PER TON FOR DIFFERENT HOT WATER FLOW RATES............... 145
FIGURE 9-5. CAPACITY COMPARISON BETWEEN SE AND HE CYCLE..................................... 146
FIGURE 9-6. COMPARISON OF CAPACITY AND COP FOR THE SINGLE AND HALF-EFFECT
ABSORPTION UNIT. ......................................................................................................... 147
FIGURE 9-7. COMPARISON OF COST OF HEAT FOR THE SINGLE AND HALF-EFFECT ABSORPTION
UNIT. .............................................................................................................................. 148
xv

xvi

LIST OF TABLES
TABLE 1-1. TYPES OF ABSORPTION UNITS ON THE MARKET*................................................... 8
TABLE 2-1. COMPARISON OF DIFFERENT SYSTEMS WITH A GAS TURBINE CHP (MELOCHE ET
AL., 1996) ........................................................................................................................ 15
TABLE 2-2. SUMMARY OF PILOT PLANT DESIGNS (SCHWEIGLER ET AL., 1999). ...................... 19
TABLE 3-1. LIMITS ON OUTLET ABSORBER SOLUTION FLOW RATE [M1A]................................ 41
TABLE 5-1. DIFFERENT TYPES OF COOLING TOWERS AVAILABLE............................................ 74
TABLE 6-1. PRICE RANGE FOR THE TUBE COST OF EACH COMPONENT (US CHILLER
MANUFACTURER, 2000). .................................................................................................. 78
TABLE 6-2. RANGE OF NUMERICAL PARAMETERS. .................................................................. 81
TABLE 7-1. BASE CASE COMPONENT SIZE AND INTERNAL SOLUTION FLOW RATE [200 F @
2000 GPM] ....................................................................................................................... 86
TABLE 7-2. COMPONENT SIZE AND INTERNAL SOLUTION FLOW RATE TO MEET 600 TONS
[200
O
F @ 2000 GPM]....................................................................................................... 97
TABLE 7-3. THE EFFECT OF INCREASING THE TUBES IN THE ABSORBER, CONDENSER, AND
EVAPORATOR WITH THE CORRESPONDING DECREASE IN GENERATOR TUBES.................... 99
TABLE 7-4. LISTS THE FINAL DESIGN FOR THE ABSORPTION CHILLER USING LOW TEMPERATURE
WASTE HEAT [600 TONS ,200
O
F @ 2000 GPM, NO SUPERHEAT 0.8 LTHX
EFFECTIVENESS]............................................................................................................. 104
TABLE 7-5. COMPARES THE CAPITAL AND OPERATING COST BETWEEN AN ABSORPTION AND
ELECTRIC CENTRIFUGAL CHILLER FOR A 20 YEAR LIFE CYCLE [~600 TONS, 0.05$/KWH,
200
O
F HOT WATER @ 2000 [GPM], 4380 HOURS] ........................................................ 113
TABLE 8-1. INITIAL HALF-EFFECT DESIGN........................................................................... 118
TABLE 8-2. OPTIMUM HALF-EFFECT DESIGN [600 TONS, 200
O
F @ 2000 GPM]................... 123
TABLE 9-1. CAPITAL COST COMPARISON BETWEEN SINGLE-EFFECT, HALF-EFFECT, AND
ELECTRIC CENTRIFUGAL CHILLER [600 TONS, 200
O
F, 2000 GPM HEAT SOURCE]............ 143
TABLE 9-2. COMPARES THE LIFE-CYCLE COST BETWEEN THE SINGLE AND HALF-EFFECT
ABSORPTION UNIT. [20 YEAR LIFE CYCLE , 600 TONS, 0.05$/KWH, 200
O
F HOT WATER @
2000 GPM, 4380 HOURS]. .............................................................................................. 149
TABLE 9-3. TUBE AND UA COMPARISON BETWEEN HALF AND SINGLE EFFECT UNIT. [600 TONS,
200
O
F HOT WATER AT 2000 GPM.] ................................................................................. 150

xvii

xviii

NOMENCLATURE

Subscripts

1,7,8,23,24,24t25l,25v,30,31,1a,1b,11,25,40,41,50,51

representative state points for the
absorption chiller

30sat

saturated steam inside the generator
11p equilibrium temperature entering the absorber
7p equilibrium temperature entering the generator
a absorber
ab absorption chiller
ab,ct absorption chiller cooling tower
a,i, a,o air inside and outside
a,eff air effective
c

condenser
e

evaporator
ec

electric chiller
ec,ct

electric chiller cooling tower
f

film
fan,nom,i

nominal fan speed
g generator
ft

finned tube
hs

heat source
hr

heat recovery
i

inside

l

liquid

o

outside

sat

saturated
st

smooth tube
s,eff

saturated effective
v

vapor
w

water

Variables

d Diamter [ft], discount rate
D Ratio of down payment to initial investment
C Capital cost [$]
C
h
break-even cost of the heat source [$]
C
min
Minimum specific heat of water [Btu/lb
m
-F]

C
s
Effective specific heat, which is a change in enthalpy divided by a change
temperature along a saturation line [Btu/lb
m
-F].
xix

C
w
Specific heat of water [Btu/lb
m
-F]
f
abs
Fraction of LC absorber tubes in half-effect cycle
f
gen
Fraction of LT generator tubes in the half-effect cycle
F
o
Rated volume flow rate of hot water [gpm]
F Volume flow rate of hot water [gpm]
g Gravitational constant [ft/s
2
]
h Enthalpy [Btu/lb
m
] / heat transfer coefficient [Btu/hr-ft
2
-F]
K Overall heat transfer coefficient [Btu/hr-F]
K
o
Rated

overall heat transfer coefficient [Btu/hr-F]
LCC Life cycle cost [$]
Le Length of tubes
LMTD Log mean temperature difference
m Mass flow rate [lb
m
/hr]
m Annual mortgage interest rate
M
s
Ratio of first year cost (maintenance, insurance) to initial investment
NC Number of tubes in a column
N
cells
Number of cells in the cooling tower
N
D
Depreciation lifetime [years]
N
e
Period of economic analysis [year]
N
L
Term of the loan [years]
N
min,
Year of mortgage payments min(N
L
or N
e
) [years]
N
min
Years which depreciation contributes min(N
D
or N
e
).
Nu Nusselt Number
NTU Heat transfer units
P
1
Ratio of life cycle fuel cost savings to first year fuel cost [~ N
e
/2]
P
2
Life cycle expenditures because of the capital investment. [~ 0.8-1.2]
P Pressure [psia]
P
elec
Electric Power consumption [hp]
P
fan
Fan Power [hp]

Pr Prandtl Number
Q Heat transfer rate [Btu/hr]
R Resistance of tubes [Btu/hr-ft
2
-F].
Re
D
Reynolds number in pipe
Re
f
Film Reynolds number
R
v
Ratio of the resale value
Subcool Subcooling at the outlet of the absorpber [
o
F]
Spill Fraction of spill out of the evaporator

T Temperature [
o
F]
t Effective income tax rate.
t Property tax based on assessed value
Tube Number of tubes
UA Overall heat transfer coefficient [Btu/hr-F]
V Ratio of assessed valuation of the system in first year to the initial investment
w Humidity ratio of the air [lbm
w
/lbm
dry,a
]
x LiBr concentration
xx

Greek

Film flow rate [lb
m
/ft-hr]
Effectiveness of heat exchanger [0.-1.]
Viscosity of the solution [lb
m
/ft-hr]
Spraying density [lb
m
/hr-ft]
efficiency of equipment
Ration of air flow to nominal air flow

xxi

xxii

CHAPTER 1

INTRODUCTION

The purpose of this project is examine the performance of a current single-effect absorption
chiller and determine what design changes are needed so that the chiller can operate using
low temperature hot water. For example, if there is a hot water source at 205
o
F with a flow
rate of 1000 gpm, what changes are needed to a machine that normally operates at 230
o
F and
1000 gpm to obtain the same capacity. The second phase of the project involves designing a
half-effect cycle to operate on low-temperature hot water. The performance characteristics
of the half-effect cycle are examined by changing the hot water flow rate, the cooling water
temperature, and the flow arrangements of the cooling and heat source flow. The half-effect
and single-effect cycles are compared to see which one is more competitive for a specified
fuel source temperature.

1

1.1 Why Absorption?

Absorption units are able to generate a cooling effect with a heat source. The heat source can
come in the form of natural gas, solar, coal, co-generation, or an industrial waste stream. A
main advantage of absorption units is their ability to utilize waste heat streams that would be
otherwise discarded.
1.2 What is Absorption?

The standard vapor compression refrigeration system is a condenser, evaporator, throttling
valve, and a compressor. Figure 1-1 is a schematic of the components and flow
arrangements for the vapor compression cycle.
chilled
water
Condenser
Evaporator
tower water
Compressor
Valve

Figure 1-1. Schematic of a centrifugal chiller.

2

Absorption refrigeration systems replace the compressor with a generator and an absorber.
The generator is a shell and tube heat exchanger, which sprays a LiBr-H
2
O solution over
heated tubes to boil off the water [refrigerant]. The refrigerant vapor then enters the
condenser. The absorber is a shell and tube heat exchanger that draws in the water vapor
from the evaporator and sprays the LiBr-H
2
O solution from the generator over cooling water
tubes to absorb the water vapor. During the absorption process a portion of the heat is
transferred to the cooling water, which comes from the cooling tower. In Figure 1-2 the
generator and absorber replace the compressor in Figure 1-1.
chilled
water
Generator
Condenser
Evaporator Absorber
(liquid spill)
LTHX
tower water

Heat Source
Figure 1-2. Schematic of the single-effect absorption chiller.

The ability for absorption to work is determined by the type of fluids that are chosen. The
LiBr-H
2
O combination uses the property that LiBr has an affinity for water. That is the
chemical potential is favorable for the absorption of water into LiBr. The water in an
3

absorption cycle acts as the refrigerant and thus is the vapor leaving the evaporator. In a
vapor compression refrigeration cycle, the water vapor would be compressed, but with
absorption the water vapor is absorbed into the LiBr-H
2
O solution and thus a liquid is
pumped instead of a vapor. The pumping of a liquid requires less energy than the
compression of a vapor. Absorption uses heat instead of electrical power to provide a cooling
effect. The high pressure LiBr solution is then sent to the generator to recover the water that
was absorbed. Figure 1-2 is a schematic of the components but Figure 1-3 is a realistic
picture of the absorption process.

Figure 1-3. Single-effect absorption chiller (US chiller Manufacturer, 2000).

4

Figure 1-3 shows a cross-section of a single-effect absorption chiller. The chiller is split into
two regions the generator-condenser at the top and the absorber-evaporator in the bottom.
The sprays are used to ensure compete wetting of the tubes for better heat transfer. There is
also a purge system that removes non-condensable gases from the system, which inhibit
absorption of the water vapor. The solution heat exchanger on the bottom left in Figure 1-3
is used to reduce the heat input at the generator.
1.3 History of large tonnage absorption units

The following summary on the history of large tonnage [>100 tons] absorption units is taken
from (Plzak, 1996).

Before the 1970s fuel oil and natural gas were inexpensive and readily available. Electricity
was also inexpensive and thus efficiency was not a concern when operating a chiller. During
this time an equal number of absorption and electric centrifugal chiller were sold.

In the 1970s, a steady increase in energy costs caused the chiller industry to examine the
efficiency of their units. The double-effect unit was developed to increase the COP
[Q
cool
/Q
in
], and thus their share in the market place using high-pressure steam increased.
The oil embargo increased the price of fossil fuels more dramatically then electric energy
prices and thus overnight killed the domestic absorption market. The absorption market has
never recovered to its pre 1973 levels.

5

The 1980s saw flat or decreasing natural gas prices and a steady increase in electrical rates.
This resulted in the growth of the newly developed single and double effect direct-fired
absorption unit in the mid to late 1980s. A direct-fired unit is one is which natural gas is
combusted in the generator instead of using the gas to create steam or hot water. The growth
was also a result of the gas and electric utility initiatives, which supported programs for gas
cooling in the summer, to keep the supply of natural gas constant throughout the year. The
phase out of CFC based refrigerants used in electric centrifugal chillers also helped boost the
direct-fired market.

The current single effect market is small but stable and is used in applications where low-
grade heat is available. The applications are chemical and manufacturing processes, small
co-generation systems, and buildings with existing low-pressure boilers. The current market
for the single effect is about 50% of the total absorption market.

1.4 Where are commercial absorption units found?

The most common applications for absorption units are in situations with high electrical cost
or a heat source of low pressure steam or hot water that is free. The types of co-generation
applications are
District heating and cooling networks
Industry / Facilities with waste heat
Schools / Hospitals

6

The single-effect is applicable in areas with sources of low-grade heat and low water costs
(IEA, 1999). The single-effect unit can be competitive with an electric chiller where electric
prices are high or the chiller operates a large portion of the year (IEA, 1999). The reason for
the advantage of longer operation time is because the operating cost is much less for the
absorption chiller and thus this could make up for the higher capital investment. Another
aspect of absorption is that twice as much water is needed for the cooling tower and thus low
water cost is also an important factor when deciding on absorption (IEA, 1999).

There is a need to develop systems for applications with temperatures in the range of 140-
176
o
F (60-80 C) (Lamp and Ziegler, 1996). The most viable alternative to cool with these
temperatures is to use the halfeffect cycle.

1.5 Who makes Absorption units?

Table 1-1 lists some of the companies that make absorption units and the characteristics of
these units.

7

Table 1-1. Types of Absorption Units on the Market*
TRANE YORK CARRIER YAZAKI SANYO
Capacity
(Tons)
100-1500 100-1500 108-608 10 100-1500
Heat Source
Hot Water

Steam

Consumption

Natural Gas

200-270
o
F

14 psig

210
o
F

15 psig

250
o
F

15 psig

167-212
o
F

94
o
F

11.4 psig
(8kg/cm2-G)
9.7lb/hr-Ton

11905
Btu/hr-Ton

Chilled Water 40-50
o
F 45
o
F 55
o
F
No cooling
Tower

40-55
o
F
48
o
F 41-54
o
F
*All data in the table is based on company website information.
8

CHAPTER 2

REVIEW OF PUBLISHED WORK ON LOW TEMPERATURE HEAT
SOURCE APPLICATIONS

2.1 Overview

The purpose of this chapter is to review the published research done on the single effect
absorption unit using a low temperature heat source. Lamp and Ziegler (1996) summarize
the idea of using low temperature heat as follows.

Since the minimum driving temperature is determined by the boundary
conditions of the evaporator, condenser, and absorber, the only way to
lower the driving temperature is to increase the heat exchanger area. The
above conclusion results in general to an economical question rather than
a technical question for achieving lower driving temperatures.

There are many published papers on the single effect unit but it is the focus of this chapter to
review low temperature applications with a single-effect lithium bromide absorption unit.
For example there are single effect units with different working fluids [Water-NH
3
] and there
9

are also solid absorption units. Chapter 2 also provides a published review of the half-effect
lithium bromide cycle.
2.2 Low temperature heat source applications using the single-effect
absorption unit

Lamp and Ziegler (1996) point out that it would be naive to obtain low firing temperatures by
only increasing the size of the generator. They use the following example,
By doubling the generator area, 0.3 tons [1 kw] of cooling can be obtained with 174
o
F [79 C]
hot water. By increasing the size of all heat exchangers by the same total area, 0.3 tons of
cooling can be obtained with 172
o
F [77.6 C] hot water.

The above example demonstrates the importance of not only looking at the generator when
designing an absorption unit to run off of a low temperature heat source. It is important to
optimize the distribution of heat exchanger area.

A single-effect lithium bromide absorption chiller computer model was validated with
experimental data (Homma et al., 1994). The capacity of the absorption chiller is 30 Tons
(105.5 kW) and is driven by waste heat in the form of hot water from a gas engine. The
absorption model consists of four main components [absorber, generator, evaporator, and
condenser]. The model of each component is based on a logmean temperature difference,
energy balance, mass balance, salt balance, and an overall heat transfer coefficient. The form
of the overall heat transfer coefficient is

10

P
o
o
F
UA UA
F
| |
=
|
\ .
[ 2-1 ]
where UA is the overall heat transfer coefficient [Btu/hr-F]
F is the volume flow rate of hot water [gpm]
P is a number determined from experiment

o
subscript o means rated value.

The temperature drop across the generator is from 190-181
o
F (88-83 C) and the evaporator
is from 55-46
o
F (13-8 C). The COP range predicted from the model is around 0.6-0.7.
Major conclusions of the paper are.

The COP and cooling capacity increase with an increase in the hot water inlet
temperature over a range of 176-201
o
F [80 94 C], thus having the exhaust gas of the
engine being as high as possible is the best design to achieve higher capacities.
The hot water flow rate can be used to control the generation of unutilized refrigerant by
controlling the COP or keeping the COP constant by changing the cooling capacity.
Decreasing the cooling water inlet temperature can increase the COP and cooling
capacity.

If the hot water inlet temperature is decreased, the capacity of the single-effect unit will also
decrease. With the addition of a compressor between the absorber and the evaporator the
same capacity can be gained with a lower generator temperature (Thornbloom and Nimmo,
1994). Thornbloom and Nimmo (1994) demonstrated this effect on a 1 ton chiller. In order
to achieve 1 ton of cooling with the conventional system, the hot water temperature needs to
be at 200
o
F (93.3 C) but with the compressor it can be lowered to 175
o
F (79.4 C). The COP
for both cycles is 0.77 [200
o
F hot water] and 0.79 [Compressor]. Without the compressor the
capacity is reduced to 0.23 tons and a COP of 0.63 with 175
o
F hot water. A disadvantage of
this cycle is the addition of the compressor to compress water vapor. For this example the
11

extra work was 0.15 hp [117 W], but for larger systems this parasitic load might not be an
advantage to reducing operating costs.

A small scale experimental single-effect lithium bromide absorption chiller was used for an
experiment in the Hot Springs of Sivas, Turkey (Kececiler,et al.,1999). A computer model
was developed based on the experimental data to determine the performance of an absorption
chiller. The availability of the geothermal heat source is too low to be used effectively in
generating electricity but the hot water source can be used to air condition at 40-50
o
F (4-10
C) (Kececiler,et al., 1999).

The results from the computer model for the optimum COP are given below.

Heat Source:
Hot water: 140
o
F (60 C)
7925 gpm (500 l/s)
99208 lb
m
/hr (12.5 kg/s)
COP 0.56
Capacity 64 Tons (225.5 kW)
Concentration 48% Generator
44% Absorber

Chilled Water 36-37
o
F (2-3 C)
Cooling Water 85-96
o
F (30-35 C)

The only way Kececiler et al., (1999) were able to utilize such a low temperature heat source
is because of the extremely high warm water flow rate. For example, 100 tons with a
temperature drop of 10
o
F would require a volumetric flow rate of approximately 300 gpm.

12

A parametric study on the input parameters shows an increase in COP for an increase in
generator or evaporator temperatures and a decrease in COP for an increase in absorber
cooling water temperature (Kececiler et al., 1999).

Compressor Turbine
Combustion
Generator
Exhaust gas
Heat recovery
boiler
District heating return
District heating supply
Air

Figure 2-1. Simple CHP using a gas turbine (Meloche et al., 1996)

Bruno et al, (1996) and Meloche et al, (1996) discuss the integration of an absorption chiller
in a combined heat and power plant [CHP]. Figure 2-1 is a schematic of a simple
combustion turbine CHP. The district heating supply can then be used to run an absorption
chiller when the heat supply is greater than the demand.

Other types of CHP systems are a reciprocating cycle or a steam turbine cycle. The
electricity produced can be used to run an electric chiller to meet the cooling load and
13

electrical load or the heat supply can be used with an absorption chiller to meet the cooling
load and hot water load. Bruno et al. (1996) examines the absorption unit as a steam
consumer, where the steam flow rate is a function of the inlet air flow to the gas turbine,
ambient air conditions, and refrigeration load. For Bruno et al. (1996) part of the cooling
load is the result of cooling the gas turbine inlet air temperature to 60
o
F [15 C], for a
maximum efficiency. Bruno et al., (1996) concludes that the absorption chiller is
economically viable with the gas turbine when the chiller is used to the cool the inlet air to
the turbine to increase generator capacity and also cover additional refrigeration needs.

Meloche et al., (1996) assumes a heat supply recovery from the heat recovery boiler at 212
o
F
[100 C] and the heat return back to the boiler at 167
o
F [75 C]. The gas turbine has an
electrical efficiency of 35%, heat efficiency of 53%, and 12% losses. Meloche et al., (1996)
examine running the gas turbine at an input of 100 units of fuel and determines the maximum
cooling using a compression chiller and an absorption unit.

The calculation is based on maximum chilled water production from all the available energy,
which is also a reason for the electricity to chillers to be large for the absorption chiller. The
absorption chillers are being supplemented with electrical chillers from the excess electrical
output. The net heating output in Table 2-1 is the amount of energy that was not useful
enough to drive a heat driven chiller (Meloche et al., 1996).

14

Table 2-1. Comparison of different systems with a gas turbine CHP (Meloche et al., 1996)
Max Cooling Electric Drive 1-Stage
HW
1-Stage
Steam
2-Stage
Steam
Fuel Input 100.00 100.00 100.00 100.00
Electricity
production
34.63 34.63 34.63 34.63
Thermal to
chillers
0.00 51.28 50.10 48.49
Electricity to
Chillers
29.85 28.14 28.25 27.77
Electricity to
auxiliary
4.78 6.50 6.38 6.86
Net electricity 0.00 0.00 0.00 0.00
Net cooling 175.14 202.97 199.07 221.10
Net heating 53.71 2.42 3.61 5.22

The electricity to auxiliary row in Table 2-1 takes into account energy needed to operate a
cooling tower fan and pumps for the cooling and chilled water. For a gas turbine CHP the 2-
stage double-effect absorption chiller provides the largest amount of cooling, but the electric
chiller is more efficient in terms of total energy output (Meloche et al., 1996). The major
conclusions of the paper by (Meloche et al., 1996) are:

There is no difference in the economics for using a heat driven chiller compared to a
compression chiller for a new CHP facility.
The difference in cooling cost between the different chillers was minimal because of the
high cost of the CHP plant and there is a small difference in efficiencies when integrated
with cogeneration in cooling mode.

The papers by Meloche et al., (1996) and Bruno et al., (1996) seem to reach different
conclusions of integrating absorption chillers with a CHP but the systems they analyzed are
different. Where Meloche is maximizing cooling output and Bruno is increasing the
15

efficiency of the gas turbine by cooling the inlet compressor air temperature. It is impossible
to conclude from these two papers if absorption is competitive with electric chillers in CHP
plants, because it is based on cooling, heating, electrical load, and ambient conditions.
2.3 Half-effect absorption unit

The half-effect absorption unit can be found in applications where the heat source
temperature is too low to be used to fire a single-effect unit (Herold, et al., 1996). The
advantage of the half-effect cycle is that the heat-input temperature can be less than the
single-effect for the same evaporator and heat rejection temperature. The disadvantage of the
half-effect cycle compared to the single-effect is the COP of the half-effect cycle is
approximately half of the single-effect, which translates that for a given capacity the heat
rejection of the half-effect cycle is larger. A half-effect cycle [50 tons [175 kW]] was built
and tested in a laboratory funded by the DOE that confirms the above conclusions from
Herold, et al., (1996).

Schweigler et. al (1996) describe a single-effect, half-effect combination unit [SE/HE]. A
single-effect machine can operate at 176
o
F (80 C), but in district heating networks the
summer temperature for hot water is in the range from 158-176
o
F (70-80 C). Thus a
combination SE/HE will allow use in district heating networks. Two advantages of the
SE/HE chiller are a reduction in heat exchanger area and it allows for a temperature glide in
the generator of 54
o
F (30 C).

16

The SE/HE chiller operates with a COP around 0.55 to 0.6 with driving heat supply / return
temperature of 194 /140
o
F (90/60 C) (Schweigler et al., 1996). The operation of the SE/HE
chiller is driven by the available heat source temperature and flow rate. If the heat source
temperature is above 176
o
F (80 C) then the SE/HE acts as a SE unit with a COP of 0.7. Any
temperatures below this the SE/HE acts as a combination SE-HE or pure HE, with the lower
bound on the COP being 0.35 (Schweigler et al., 1996).

The control of the SE/HE at part load conditions can be met by decreasing the hot water flow
rate or temperature. An important conclusion is that part load control by temperature reduces
the COP, but controlling part loadwith mass flow rate, the COP slightly increases until 20%
part load.(Schweigler et al., 1999). Part load below 20% the SE/HE operates as a half effect
cycle and thus the COP drops to around 0.35-0.4.
Figure 2-2 displays the single-effect / half-effect combination. In Figure 2-2 the evaporator
(E0), absorber (A0), condenser (C2), and generator (G21) are for the SE subcycle and the E0,
A0, C2, G1, and G22 are for the HE cycle. Even though it looks like the SE/HE has three
generators there is only two, where the G2 generator is able to act like two by how the flow is
arranged.

17

Figure 2-2. Flow chart of the SE/HE chiller (Schweigler et al., 1999).

The increase in COP at part load is gained by reducing the fraction of heat to the HE
generators [G22, G1] more than the SE generator [G21] (Schweigler et al., 1999). This
process is done by splitting the flow after it leaves the G21 generator and thus only a portion
of the mass flow rate enters into G1 and G22 (Schweigler et al., 1999). This process can also
control the temperature returning to the district-heating network. Mixing the flow from the
G22 generator with the by-pass flow can maintain the desired return design point
temperature, which is usually around 140
o
F (60 C) (Schweigler et al., 1999).

Three pilot plant designs were implemented that have the SE/HE design. The Berlin and
Dsseldorf pilot plants were designed for use in a district heating network and the other is
used for co-generation in the Munich airport. Table 2-2 summarizes the operating
parameters and size of the units.
18

Table 2-2. Summary of pilot plant designs (Schweigler et al., 1999).
District Heating Co-Generation

Installation Technical University,
Berlin
Power Plant
Lausward,
Dsseldorf
Mnchen Airport
Manufacturer Entropie
GmbH,
Erding,Germany
GEA
Luftkhler
GmbH,
Herne,Germany
Entropie
GmbH,
Erding,Germany
Design Point
Driving Heat
o
F (C) 203/150 (95/65) 185/140 (85/60) 203/140 (95/60)
Cooling Water
o
F (C) 81/95 (27/35)
Chilled Water
o
F (C) 54/43 (12/6)
Cooling Capacity
Tons (kW)
114 (400) 85 (300) 710 (2500)
COP 0.62 0.58 0.65
Dimensions ft
LxWxH (m)
14.7x6.9x6.9
(4.5x2.1x2.1)
18x6.9x9.8
(5.5x2.1x3)
19.7x9.8x16.4
(6x3x5)
Weight
tons (metric tons)

14.3 (13)

16.5 (15)

55.1 (50)

An important aspect of the SE/HE design is that the generators used in the SE/HE unit are
falling film generators with multi-pass arrangement to ensure good heat transfer and large
temperature glides. The units are also designed high and narrow in a rectangular profile to
avoid the unfavorable large solution flow rates (Schweigler et al., 1999).

The main conclusions from the pilot plant test are (Schweigler et al., 1999).
Operating expenses matched those with the theoretical performance predictions.
COP increased during part load conditions while the temperature glide in the hot water
remained constant.
An increase in cooling capacity was achieved per unit of hot water mass flow compared
with a standard single effect unit

19

Ma et al., (1996) examined a 100 ton (350 kW) half-effect absorption chiller driven by low
temperature waste heat from a combined heat and power (CHP) plant in Beijing, China. The
halfeffect chiller is able to operate at a hot water temperature of 187
o
F (86 C) to produce
49
o
F (9 C) chilled water using 87
o
F (32 C) cooling water at a COP of 0.4(Ma et al., 1996). A
single-effect chiller can not produce the specified chilled water with the given heat source
temperature so a half-effect device must be used.

Erickson D. (1995) describes a half-effect cycle using (NH
3
-H
2
O), which is referred to as the
vapor exchange cycle. The main advantage of this cycle over the LiBr-H2O unit, is that the
evaporator temperature can go below 32
o
F [0 C] because ammonia is the working fluid in the
evaporator. The state of Alaska, the Alaska Science and Technology Foundation, and Alaska
Energy Authority [AEA], helped fund the project. The capital cost for the 10 ton unit is
about 7,000 $/ton but with 0.21 $/kWh the higher capital cost is worth the initial investment
to reduce operating cost. The COP for half-effect cycle is about 0.306, which was determined
from measuring the heat input and ice making output. The heat source comes from the jacket
and exhaust of a diesel generator.

20

2.4 Summary

From examining the published literature the single-effect unit finds applications in co-
generation, geothermal energy, and combined heating a cooling plants. The best method for
designing single-effect units too operate on a low temperature heat source is to optimize the
heat exchanger area of all components.

The half-effect unit can operate using a lower temperature heat source compared to the
single-effect, but this ability comes at a cost, which is a lower cooling COP. No published
literature was found that provides a detailed study of the complete system of absorption unit
and cooling tower.

21

2 2

CHAPTER 3

SINGLE-EFFECT ABSORPTION MODEL DEVELOPMENT

3.1 Overview

Figure 3-1 is a schematic of the flow arrangement for a single-effect absorption unit. The
main components of the absorption unit are; generator, condenser, evaporator, absorber, and
low temperature heat exchanger [LTHX]. The brine or LiBr-H
2
O solution is pumped from
the absorber to the generator where the water is boiled off. The heat source is passed in a
counter-flow arrangement through the generator to boil off water vapor from the LiBr-H
2
O
solution. The pressure of the system is about 1 psia at the generator and 0.1 psia at the
absorber. This low pressure or vacuum system allows water to be used as a refrigerant.

23

chilled
water
Generator
Condenser
Evaporator Absorber
(liquid spill)
LTHX
tower water
10
1b
1a
23
8
7
2
24t
25
24
30
11
51 50
42
40
31

Heat Source
Figure 3-1. Schematic of the Single Effect Absorption Unit.

A cooling water loop is needed to condense the water vapor boiled off from the generator and
to aid in the absorption of water vapor back into the LiBr-H
2
O solution. This cooling water
is passed first through the absorber and then the condenser. The evaporator takes in low-
pressure cold water and produces a cooling effect by evaporating the water and passing it to
the absorber.

A critical point in the system is the solution leaving the LTHX. The solution consists of a
concentrated solution in LiBr at the lowest temperature in the cycle and thus a process called
crystallization can occur.. Crystallization is the formation of salt crystals due to a low
temperature high LiBr-H
2
O solution. Crystallization must be avoided because the formation
of a wet solid (slush) in the piping network over time could form a solid and block the flow
24

(Herold et al., 1996). Crystallization can be avoided by adding heat to the system in that area
or diluting the solution with water from the evaporator (Herold et al., 1996)

A recirculation pump is used for the evaporator and absorber to ensure complete wetting of
the tubes. The numbers in Figure 3-1 represent the states points used in the computer model.
For example, T30 is the temperature of the entering hot water.
3.2 Model Input Parameters
3.2.1 Generator
Steam

T30 Temperature of Steam [F]
P30 Pressure of Steam [F]
H_ig Inside heat transfer coefficient [Btu/hr-ft
2
-F]
Tube# Total Number of Tubes
NCG Number of Tubes in column
D_g Inner and Outer Diameter of Tubes [in]
Le_g Length of Tubes [ft]

Hot Water

T30 Temperature of Water [F]
Vol30 Volume flow rate of Hot Water [gpm]
NCG Number of Tubes in column

3.2.2 Condenser
UA_cond [Btu/hr-F]
Or
Tube# Number of Tubes
D_tube Inner and Outer Diameter of Tubes [in]
Le_tube Length of Tubes [ft]

m41 Mass Flow Rate of Cooling Water [lbm/hr]
25

3.2.3 Evaporator
UA_evap [Btu/hr-F]
Or
L_tube Length of Tubes [ft]

NCE Number of Tubes in a column
m50 Mass Flow Rate of Chilled Water [lbm/hr]
T51 Temperature of Delivered Chilled Water [F]
Spill Percent of liquid that is spilled directly to absorber [%]

3.2.4 Absorber
T40 Temperature of Entering Cooling Water [F]

NCA Number of Tubes in a column

3.2.5 Low-Temperature Heat Exchanger
_lthx Effectiveness of Low-Temperature Heat Exchanger
3.3 Assumptions

3.3.1 Overall
Steady-state.
The pressure in the evaporator is equal to the absorber and likewise for the generator and
condenser.
No pressure drop or heat transfer in the piping network.
All components are modeled as shell and tube heat exchangers. The log mean
temperature difference is used in the generator/absorber and effectiveness-NTU method
for condenser/evaporator.
The throttling device isoenthalpic.
The solution pump is adiabatic.
26

3.3.2 Generator
Equilibrium condition at the outlet of the generator point [8].
No carry over of salt into the refrigerant.
The steam leaves as saturated liquid.
The outlet temperature of the refrigerant is the average of the equilibrium entering
temperature and the outlet temperature. T[23]=(T7p+T[8])/2

3.3.3 Condenser
The outlet of the condenser is saturated liquid at the generator pressure.
3.3.4 Evaporator
The outlet of the evaporator is saturated vapor.
A 4% spill of liquid from the evaporator goes directly to the absorber without
evaporating.

3.3.5 Absorber
The outlet absorber solution is sub-cooled.
The absorber spray solution is an empirical function of capacity and inlet cooling
temperature.

3.4 Model development for steam fired absorption unit
3.4.1 Overall

The four components (absorber, generator, condenser, and evaporator) are modeled based on
energy balances for the internal and external streams, a heat transfer rate equation between
the streams, salt balances, and mass balances. The overall heat transfer coefficient [UA] is
determined from empirical relationships for the inside and outside heat transfer coefficient.
The properties of enthalpy, pressure, temperature, concentration are determined from
property relations. The mass and salt balance for the generator is of the form, where the
numbers are based on Figure 3-1.
27

[ 3-1 ]
8 23 7
m m m + =
[ 3-2 ]
8 8 7 7
x m x m =

The mass and salt concentration leaving the absorber is set equal to the mass entering the
generator. The high and low pressures are determined from the saturated pressure of the
water vapor (Electrical Research Association, 1967 Steam Tables).
3.4.2 Generator

Energy balances on the internal solution and heat source in the generator are

8 8 23 23 7 7
h m h m Q h m
g
+ = + [ 3-3 ]
[ 3-4 ]
30 30 30 31 g
m h Q m h =
The heat transfer rate is calculated by
[ 3-5 ]
g g g
LMTD UA Q =
where the LMTD is defined as

|
|
.
|
\
|
=
sat
sat p
p
g
T T
T T
T T
LMTD
30 8
30 7
7 8
ln
[ 3-6 ]

where T
7p
is the equilibrium temperature of the solution entering the generator at the
generator pressure and solution salt concentration [F].
T
30sat
is the saturation temperature of the steam at the given pressure [F].

28

Since the generator is steamed fired it does not matter if it is counter or parallel flow
arrangement because the heat transfer is taking place at a constant steam saturation
temperature.

The overall heat transfer coefficient is calculated using a resistance type equation.

1
1 1 1 1
g
ig ig og og g ig g
UA
h A h A R A Tube
=
| |
+ +
|
|
\ .
[ 3-7 ]

where h
ig
is the inside heat transfer coefficient in the generator [Btu/hr-ft
2
-F].
h
og
is the outside heat transfer coefficient in the generator [Btu/hr-ft
2
-F].
R
g
is the resistance of the metal [Btu/hr-ft
2
-F].
Tube
g
is the number of tubes in the generator.

The inside heat transfer coefficient for steam is assumed to be a constant at 1600 [Btu/hr-ft
2
-
F] (US Chiller Manufacturer, 2000). The outside heat transfer coefficient is based on an
empirical relationship derived from data provided by a US chiller manufacturer for their
single-effect unit. Figure 3-2 displays a graph of the relationship.
29

R2 = 0.9451
0
500
1000
1500
2000
2500
0 10 20 30 40 50
Film Reynolds Number
N
u
s
s
e
l
t

N
u
m
b
e
r
Nu_data
Model

Figure 3-2. Empirical relationship between outside Nusselt number and film Reynolds
number for the generator.

The data in Figure 3-2 are based on numerous experiments with different steam pressures /
saturation temperatures and absorber inlet cooling water temperatures. A log fit was used to
approximate the data with an R
2
of .95. Equation 3-8 is the log fit for the Nusselt number
[ 3-8 ] 1087.3ln(Re ) 2189.4
g fg
Nu =

where Re
fg
is the film Reynolds number

g
4
Re
g
fg
f
= [ 3-9 ]
where
g
is the film flow rate [lb
m
/ft-hr]
fg
is the viscosity of the salt solution [lb
m
/ft-hr]

The film flow rate is defined as
30

7
2
g
tube
m
le NC
=

[ 3-10 ]
where NC is the number of tubes in a row.

The outside heat transfer coefficient is calculated from

g og
og
fg
Nu D
h
k
= [ 3-11 ]
where k
f
is the conductivity of the salt solution [Btu/hr-ft-F].
D
o
is the outside diameter of the tubes [ft].

The model restricts the outside Nusselt number to be below 2000 so as not extrapolate past
the data.

3.4.3 Absorber

Energy balances on the internal solution and the cooling water in the absorber are
[ 3-12 ]
11 11 25 25 1 1 1 1 a a a b
m h m h Q m h m h + = +
b
a
[ 3-13 ]
40 40 41 41 a
m h Q m h + =
The heat transfer rate is calculated by
[ 3-14 ]
a a
Q UA LMTD =
where LMTD is defined as

11 41 1 40
11 41
1 40
( ) (
ln
p
a
p
T T T T
LMTD
T T
T T

=
| |
|
\ .
)
[ 3-15 ]

31

where T
11p
is the equilibrium temperature of the salt solution at x
11
concentration [F].

The LMTD is based on a counter flow heat exchanger arrangement.
The inside heat transfer coefficient is based on the Dittus-Boelter correlation (Incropera &
DeWitt, 1996).

.8 .4
.023Re Pr
ia ia
ia a a
wa
h d
Nu
k
= = [ 3-16 ]
where k
wa
is the conductivity of the cooling water at an average inlet and outlet conditions
[Btu/hr-ft-hr]
d
ia
is the inner diameter of the pipe [ft].

The Gnielinski correlation would increase the accuracy of the inside heat transfer coefficienct
but to be consistent with the data provided by the US chiller manufacturer the Dittus-Boelter
correlation is used.
The outside heat transfer coefficient is also based on a empirical relationship of the form
(Cosenza and Vliet, 1990)
(Re )
oa fa b
oa fa
fa
h d
Nu a
k
= = [ 3-17 ]
where k
f
is the conductivity evaluated at a mean absorber temperature and concentration
[Btu/hr-ft-hr]
d
fa
is the average laminar film thickness around the tube bundle [ft]

1
3
2
3
fa a
fa
fa
d
g
| |
= |
|
\ .
[ 3-18 ]
where
fa
is the viscosity of the solution [lb
m
/hr-ft].
a
is the film flow rate per tube column per length of tube [lb
m
/hr-ft].

The overall UA is calculated from
32

1
1 1 1 1
a
ia ia oa oa a ia a
UA
h A h A R A Tube
=
| |
+ +
|
\ .
[ 3-19 ]

Figure 3-3 demonstrates the correlation used by Nusselt number in equations 3-20 and 3-21.
The symbols represent data points from the US chiller manufacturer and the numbers 85, and
65represent the inlet cooling temperature in
o
F to the absorber. Equation 3-20 represents the
data for a cooling inlet temperature of 75
o
F.
[ 3-20 ]
1.6258
85
.0048Re
fa
Nu =
[ 3-21 ]
1.6428
65
.0036Re
fa
Nu =
R
2
85= 0.9357
R
2
65= 0.9679
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 10 20 30 40 50
Film Reynolds Number
N
u
_
a
b
s
Nu85
Nu75
Nu65
Model 85-75
Model-65

Figure 3-3. Nusselt number as a function film Reynolds number.

An important condition depicted in Figure 3-3 is the decrease in the Nusselt number for a
decrease in absorber inlet cooling temperature. The most likely reason for this result is that
33

the mass flow rate decreases with a decrease in inlet cooling water. The absorption model
prevents extrapolation beyond the data in Figure 3-3. The upper limit on the film Reynolds
number is 31, which was derived from taking an average of film Reynolds numbers in the
data set. This upper limit will show itself in chapter 8 when examining the effect on capacity
by changing the internal solution flow rate.
R
2
= 0.7625
0
2
4
6
8
10
12
14
16
0 20 40 60 8
(T40+T41)/2-T25 [F]
S
u
b
c
o
o
l

[
F
]
0

Figure 3-4. Sub-cooling as a function of evaporator and absorber inlet cooling temperature.

Sub-cooling takes into account the fact that if the equilibrium concentration leaving the
absorber was calculated from the leaving temperature, the concentration would be to low.
Figure 3-4 shows the trend of sub-cooling in the data and equation 3-22 is the empirical
relationship derived from the data (US chiller manufacturer, 2000).
904 . 19
2
2569 .
25
41 40
+ |
.
|
\
|

+
= T
T T
Subcool [ 3-22 ]
The linear fit has a low R
2
value, but the exact value of sub-cooling is not important. More
important than the value is the trend of sub-cooling to acquire a rough estimate. The value of
sub-cooling is used when solving for the concentration at the exiting solution of the absorber.
34

[ 3-23 ] ) , (
25 1 1
T Subcool T xeq x + =

where xeq is a function from ASHRAE Fundamentals (1989), based on LiBr properties.

3.4.4 Condenser
Energy balances on the internal solution and the cooling water in the condenser are
[ 3-24 ]
23 23 24 24 c
m h Q m h =
[ 3-25 ]
41 41 42 42 c
m h Q m h + =
The effectiveness or ratio of heat transfer rate to maximum heat transfer rate is calculated by

42 41
24 41
c
T T
T T
[ 3-26 ]
where effectiveness is defined by Incropera & DeWitt, (1996).
[ 3-27 ] 1
c
Ntu
c
e

=

Equation 3-27 is for all types of heat exchangers [concentric tube, shell and tube, and cross
flow with multiple passes] that go through a phase change. The Ntu is defined by

min
c
c
UA
Ntu
C
= [ 3-28 ]

where C
min
is the mass flow rate of the cooling water times the heat capacity [Btu/hr-F].

The C
min
is based on the cooling water because the capacity is infinite for a substance that
goes through a phase change. The inside heat transfer coefficient is based on the Gnielinski
correlation (Incropera & DeWitt, 1996).
35

( )
1
2 3
2
( 8) (Re 1000) Pr
1 12.7 8 (Pr 1)
ic wc wc c
ic
ic ic
ic
Nu k k f
h
d d
f
= =
+
[ 3-29 ]
The Gnielinski correlation was used because it is claimed to have a higher level of accuracy
over the Dittus-Boelter correlation (Incropera & DeWitt, 1996). The Gnielinski correlation is
applicable over a wider range of Reynold and Prandtl numbers than the Dittus-Boelter
equation. The outside heat transfer coefficient is based on correlation for laminar film
condensation on horizontal tubes (Rohsenow, et.al. 1998).

1
2
3
1 3
2
( ) 3600
1.51Re
lc lc vc
oc lc fc
lc
g
h k

| |
=
|
\ .
[ 3-30 ]

where is the average outside heat transfer for an entire column in a tube bundle
oc
h
[Btu/hr-ft
2
-F].
Re
fc
is the film Reynolds number in a tube bundle.

The published correlation under predicts the outside heat transfer coefficient and thus a factor
of 2.5 is used in equation 3-31. The factor of 2.5 was determined from calibration with data
from a US chiller manufacturer.

1
2
3
1 3
2
( ) 3600
2.5 1.51Re
lc lc vc
oc lc fc
lc
g
h k

| |
=
|
\ .
[ 3-31 ]

The most likely reason for the under prediction is that the correlation is for laminar flow and
with a film Reynolds number on the order of 100, there could be wavy effects that equation
3-30 does not take into account, which would increase the heat transfer coefficient.

The overall UA for the condenser is based on
36

1
1 1 1 1
c
ic ic oc oc c ic c
UA
h A h A R A Tube
=
| |
+ +
|
\ .
[ 3-32 ]

3.4.5 Evaporator

Energy balances on the internal solution and the chilled water in the evaporator are
[ 3-33 ]
24 24 25 25 25 25 t t e l l v
m h Q m h m h + = +
v
[ 3-34 ]
50 50 51 51 e
m h Q m h =

The l and v represent the liquid and vapor that are spilled into the absorber. The idea of the
liquid spill is that cold water that could be used to take in heat by evaporating is spilled over
into the absorber and thus the machine has lost some cooling ability. Spill can occur because
the evaporator is on top of the absorber [refer to Figure 1-3]. The fraction of m
25l
and m
25v
is
calculated by

[ 3-35 ]
25 25 l
m m Spill =
[ 3-36 ]
25 25
(1 )
v
m m Spill =

where the spill fraction is usually around 3-5%.
The effectiveness or ratio of heat transfer rate to maximum heat transfer rate is calculated by

51 50
25 50
e
T T
T T
[ 3-37 ]
where effectiveness is defined by Incropera & DeWitt, (1996).
37

[ 3-38 ] 1
e
Ntu
e
e

=

The Ntu is defined by the UA

min
e
e
UA
Ntu
C
= [ 3-39 ]

where C
min
is the mass flow rate of the chilled water times the heat capacity [Btu/hr-F].

The inside heat transfer coefficient is obtained from a correlation provided by a tubing
manufacturer for double cut tubes. The double cut tubes represent fins on the inside and
outside.
[ 3-40 ]
1
Re Pr
b n
ie o De ie
Nu b =

The constants b
o
, b
1
, and n are fitted parameters provided by the tubing manufacturer.

ie ie
ie
we
Nu D
h
k
= [ 3-41]
The outside heat transfer coefficient is based on laminar evaporating films (Rohsenow, et.al.
1998).

1
2 2 3
3
1 3
2
3600
1.1006Re
le le
oe fe
le
g k
h

| |
=
|
\ .
[ 3-42 ]
Re
fe
<20-30
The published correlation under predicts the outside heat transfer coefficient based on
experimental data and thus a factor of 2.75 is used in equation 3-42. The factor of 2.75 was
determined from calibration with data from a US chiller manufacturer.

38

1
2 2 3
3
1 3
2
3600
2.75*1.1006Re
le le
oe fe
le
g k
h

| |
=

\ .
|
[ 3-43 ]

The constant factor of 2.75 in equation 3-43 causes the heat transfer coefficient to be under
predicted at full load and over predicted at part load. One possible reason why equation 3-42
under predicts is that the Reynolds number is calculated from the refrigerant flow rate but
since the evaporator is using recirculation, the flow rate is higher, thus a larger Reynolds
number. A larger Reynolds number could cause turbulence or a wavy region, which increases
the heat transfer. The tubes are finned on the outside so the factor 2.75 could be thought of
as increasing the surface area due to calculating the outside area based on smooth tube
geometry. No data were available for fin pitch or height. Since the evaporator is using re-
circulation the UA for the evaporator should be a constant, and thus the data demonstrates a
relatively constant UA for full and part load conditions.
The UA for the evaporator is determined by

1
1 1 1 1
e
ie ie oe oe e ie e
UA
h A h A R A Tube
=
| |
+ +
|
\ .
[ 3-44 ]

3.5 Control Strategy

In order to model a single-effect absorption unit it is important to have a control strategy.
Since the current design employs recirculation for the absorber, two control strategies are
needed. The reason for these control strategies is so the computer model can predict what the
internal solution and re-circulation flow rate should be for a given capacity and temperature
39

margin. The T
margin
is the temperature difference between the low temperature concentrated
solution leaving the LTHX and the crystallization point at that concentration
[ 3-45 ]
ation crystilliz LTHX in m
T T T =
, 10 arg

The capacity of the unit determines what the leaving flow rate out of the absorber should be
[m1a] and the temperature margin regulates what needs to be re-circulated from the absorber
[m1b].

Figure 3-5 shows a plot of relationship between capacity and absorber solution flow rate.
The numbers 85, 75, and 65 represent the inlet cooling water temperature in
o
F. The data
points in Figure 3-5 represent calculated data from a US chiller manufacturer. Recall m1a is
the solution flow rate leaving the absorber and entering the generator.

R
2
85 = 0.9271
R
2
75= 0.9884
R
2
65= 0.9679
0
20000
40000
60000
80000
100000
120000
140000
160000
180000
200000
0 100 200 300 400 500 600 700 800
Capacity [Tons]
S
o
l
u
t
i
o
n

m
a
s
s

f
l
o
w

r
a
t
e

l
e
a
v
i
n
g

a
b
s
o
r
b
e
r

[
m
1
a
]

[
l
b
m
/
h
r
]

m1a 85F
m1a 75F
m1a 65F
Model-85
Model-75
Model-65

Figure 3-5. Absorber outlet solution flow rate as a function of capacity and temperature.

40

The three exponential relationships for the internal solution flow rate were chosen based on a
high R
2
value and reproducibility of the data. The following equations are:

[ 3-46 ]
.0021
85
1 38155
Capacity
m a e =
[ 3-47 ]
.002
75
1 34788
Capacity
m a e =
[ 3-48 ]
.0018
65
1 35137
Capacity
m a e =

The absorption model limits the mass flow rates above a specified capacity based on what the
inlet cooling water temperature is. TABLE 3-1 specifies the limits of equations 3-46, 47, and
48. Also note in Figure 3-5 the dependence on the absorber inlet cooling temperature. The
reason for the temperature dependence is that the lower inlet cooling absorber temperature is
able to remove heat more effectively and thus a lower solution flow rate is needed to obtain
the same capacity.
The advantage of this control strategy is that the input [capacity] is a parameter that is usually
known and measured with high accuracy.

Table 3-1. Limits on outlet absorber solution flow rate [m1a].
85
o
F

75
o
F 65
o
F
Capacity [Tons] 620 680 700
m1a [lbm/hr] 140,000 135,000 125,000

The limits were found by taking an average of the m1a data at full load for each of the inlet
cooling temperatures.

41

3.6 Calibration

The calibration phase consisted of supplying the model with the absorber to generator flow
rate and recirculation flow rate in the absorber. The model then calculated the T
margin
,
capacity, temperatures, concentration, etc. One argument that was made was to calibrate the
model with concentration instead of mass flow rates; the reason being is that concentration
was actually measured where as the mass flow rates were calculated by a data reduction of
the measured parameters. The main reason against this technique is that a small change in
concentration results in a large change in mass flow rates. Thus by forcing the concentration,
one has the potential to produce large errors in flow rates. If the mass flow rate is forced the
potential to produce errors in concentration is low.

The next step was to implement both control strategies, [T
margin
and the empirical
relationships from Figure 3-5]. The model then predicted performance based on the two
control strategies and the above model development equations. If the predicted values
followed the trend of the data [mass flow, concentration, temperature] and were within 5-
10%, then the calibration was done. Part load conditions required the T
margin
to be slightly
adjusted from the control set point to fit the measured data of concentration and mass flow
rate.

The numerical values for the outside heat transfer coefficients for the condenser and
evaporator were also determined once the proper control strategy was fixed. Figures 3-6 to 3-
42

18 show the relationship between measured data and model predictions for capacity, COP,
UA, concentration, and mass flow rates. In each of the graphs a 45
o
degree line is used to
shows deviations from the model and data set.

One important clarification is that when the term data is used for capacity, COP, UAs and
flow rates it really means calculated values from a chiller manufacturer, the only measured
data are concentration and temperature. The manufacture has provided a proprietary analyses
program to estimate values of capacity, COP, UAs, and flow rates based on the measured
temperatures and concentration. It is these estimates to which the model developed in this
thesis is compared

0
100
200
300
400
500
600
700
800
900
0 100 200 300 400 500 600 700 800 900
Capacity Data [Tons]
C
a
p
a
c
i
t
y

m
o
d
e
l

[
T
o
n
s
]

Figure 3-6. Comparison between model and data for capacity

43

0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85
COP data
C
O
P

m
o
d
e
l

100
75
50
25
% Load

Figure 3-7. Comparison between model and data for COP.

The capacity predictions are in good agreement but the COP shows the model is not able to
predict COP at 25% of full load. The main reason for this is that the model was able to
predict the COP at 25 % full load, but it also predicted negative re-circulation rates. When a
lower limit of zero was placed on the re-circulation flow rate the COP estimates at 25% of
full load are over-predicted. The data set given by the US chiller manufacturer also
contained negative re-circulation rates.
44

0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.00 0.05 0.10 0.15 0.20 0.25 0.30
UA_g data [MBtu/hr-F]
U
A
_
g

m
o
d
e
l

[
M
B
t
u
/
h
r
-
F
]

Figure 3-8. Comparison between model and data for UA generator.
0.00
0.20
0.40
0.60
0.80
0.00 0.20 0.40 0.60 0.80
UA_a data [MBtu/hr-F]
U
A
_
a

m
o
d
e
l

[
M
B
t
u
/
h
r
-
F
]

Figure 3-9. Comparison between model and data for UA absorber

The model predictions for the UA of the absorber and generator are in good agreement but
the condenser and evaporator, show considerable scatter of data.
45
0.50
0.60
0.70
0.80
0.50 0.60 0.70 0.80
UA_c data [MBtu/hr-F]
U
A
_
c

m
o
d
e
l

[
M
B
t
u
/
h
r
-
F
]

Figure 3-10. Comparison between model and data for UA condenser.
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
1 1.2 1.4 1.6 1.8 2
UA_e data [MBtu/hr-F]
U
A
_
e

m
o
d
e
l

[
M
B
t
u
/
h
r
-
F
]

Figure 3-11. Comparison between model and data for UA evaporator.

As will be shown later, the UA of the condenser and evaporator do not matter because each
component is oversized and a 10-15 % error will not effect the system performance
significantly.
46

The concentration and mass flow rate graphs demonstrates the models ability to predict
performance.
0
50000
100000
150000
200000
0 50000 100000 150000 200000 250000
m1a data [lbm/hr]
m
1
a

m
o
d
e
l

[
l
b
m
/
h
r
]
m1a= mass flow rate leaving absorber
0
20000
40000
60000
80000
100000
-20000 0 20000 40000 60000 80000 100000
m1b data [lbm/hr]
m
1
b

m
o
d
e
l

[
l
b
m
/
h
r
]

m1b=re-circulation flow rate
Figure 3-12. Comparison between model
and data for mass flow rate leaving the
absorber.
Figure 3-13. Comparison between model and
data for re-circulation flow rate.

In Figure 3-13 some of the data shows a negative re-circulation flow rate but the model re-
circulation flow rate was limited to positive values and thus over predicts since it can not go
below zero. These negative re-circulation flow rates are the cause for the higher COP at
some part load conditions. Notice that even though there is a large error in re-circulation it
still does not have a large effect on capacity. Figure 3-12 shows good agreement with data.
Figure 3-14 and Figure 3-15 demonstrates good agreement with the data.

47

0
5000
10000
15000
0 5000 10000 15000
m[23] data [lbm/hr]
m
[
2
3
]

m
o
d
e
l

[
l
b
m
/
h
r
]
m23= refrigerant flow rate
and data for refrigerant flow rate.
0
2000
4000
6000
8000
10000
12000
14000
16000
0 2000 4000 6000 8000 10000 12000 14000 16000
m[30] data [lbm/hr]
m
[
3
0
]

m
o
d
e
l

[
l
b
m
/
h
r
]
m30 = steam flow rate
and data for steam flow rate.

56
57
58
59
60
61
62
63
64
56 57 58 59 60 61 62 63 64
x[1] data [%]
x
[
1
]

m
o
d
e
l

[
%
]
Concentration Leaving Absorber
and data for concentration leaving absorber.
60
61
62
63
64
65
66
60 61 62 63 64 65 66
x[8] data [%]
x
[
8
]

m
o
d
e
l

[
%
]
Concentration Leaving Generator
and data for concentration leaving generator.
48

59
60
61
62
63
64
65
66
59 60 61 62 63 64 65 66
x[11] data [%]
x
[
1
1
]

m
o
d
e

[
%
]
l
Concentration Entering Absorber

Figure 3-18. Comparison between model and data for concentration entering absorber.

Figure 3-16 - Figure 3-18 compares the concentration at the generator and absorber. The
graphs show small differences between model and data.
The symbols in Figure 3-19 and Figure 3-20 represent the data and the lines are the
absorption model, T
51
is outlet temperature of the chilled water. Figure 3-19 compares the
model predictions of capacity for various inlet absorber cooling water temperatures as a
function of part load.
49

Figure 3-19. Capacity as a function of load for different inlet absorber cooling temperatures.

The above graph shows that the model is able to follow the trend of the data. Two important
characteristics of this plot that will be seen later is the linear relationship of capacity with
load and the increase in capacity as the cooling water is decreased for a given load.

Figure 3-20. COP as a function of load for different inlet absorber cooling temperature.
50

Even though Figure 3-20 is an expanded scale, there is a distinct deviation from data at 25 %
part load condition. There are small differences at 100% due to 1 or 2 data points but for the
most part the model is able to predict the data for a wide range of part load conditions. Again
the trend of the COP for part load conditions is very important and will be seen in future
chapters. COP is almost liner with load because as the cooling capacity is decreased so is the
heat source.

3.7 Model Development: Hot Water Fired Generator

The hot water fired generator is a counter flow 2-pass shell and tube heat exchanger. The
overall heat transfer coefficient [UA] is calculated from an inside and outside heat transfer
coefficient. The inside coefficient is a standard empirical relationship for water and the
outside coefficient is based on a paper by Wang, et al. (1999) for a falling film generator.

3.7.1 Difference between steam fired and hot water fired generator

The main difference between the two generators is that a steam fired generator pulls in the
amount of steam needed for a capacity by condensation and thus the mass flow rate of the
steam is not an input to the model. Instead of specifying the steam flow rate, an assumption is
used that the steam leaving the generator is saturated water and thus the model calculates the
steam flow rate. For the hot water system the mass flow rate is an input and the leaving
water temperature is determined from the model.

51

Another main difference between the absorption system with hot water and steam is that the
recirculation flow to the absorber is set to zero and thus the control strategy for temperature
margin is no longer needed. The recirculation was removed because future designs will
incorporate an absorber that will be taller and narrower than the current design. A taller
design will ensure that the tubes are fully wetted.

3.7.2 Inside Heat Transfer Coefficient

The Gnielinski correlation is used for the generator inside heat transfer coefficient (Incropera
& DeWitt, 1996).

( )
1
2 3
2
8 1000
1 12 7 8 1
ig ig Dg ig
ig
wg
ig
h d ( f )(Re ) Pr
Nu
k
. f (Pr )
= =
+
[ 3-49 ]
where h
i,g
is the inside heat transfer coefficient [Btu/hr-ft
2
-F]
d
ig
is the inner pipe diameter [ft]
k
wg
is the conductivity of the water [Btu/hr-ft-F]
Re
Dg
is the Reynolds number for pipe flow.
Pr
ig
is the Prandtl number
Nu
ig
is the Nusselt number
f is empirical relationship for smooth tubes based on the Moody diagram[Incropera &
DeWitt, 1996].

2
79 1 64
Dg
f (. ln Re . )
= [ 3-50 ]

3.7.3 Outside Heat Transfer Coefficient

The outside generator heat transfer coefficient h
og
was developed by integrating three graphs
from (Wang et al., 1999). The three graphs are falling film outside heat transfer coefficient
52

versus spraying density and pressure, and leaving concentration as a function of spraying
density for a falling film generator using hot water. The spraying density [] is the mass of
brine solution per unit length of tube and time. Figure 3-21 and Figure 3-22 displays the
relationship of the outside heat transfer coefficient and exiting concentration as a function of
pressure and spraying density. Figure 3-23 displays the change in outside heat transfer
coefficient as a function of spraying density for smooth and finned tubes.

Figure 3-21. Change in outside heat
transfer coefficient as a function of
generator equilibrium pressure (Wang et
al., 1999).
Figure 3-22. Change in exiting
concentration as a function of spraying
density (Wang et al., 1999).
53

Figure 3-23. Change in outside heat transfer coefficient as a function of generator spraying
density Wang et al., (1999).

The following is a summary of the experimental set-up and parameters (Wang et al., 1999)
It is an open cycle system
The generator consists of eight rows of horizontal tubes
The tubes are approximately .4 m / 1 ft in length and 16mm / .6 inch inner diameter.
The inlet LiBr solution is between 55-56.7 %.
The hot water at the inlet is between 80-96
o
C / 176-204
o
F
The spraying density range was 85-510 [kg/hr-m] / 33-370 [lbm/hr-ft]

The geometry of the finned tube is a fin height of 1.4 mm / 0.055 in , fin width of 0.3 mm /
0.012 in, and the distance between fins is 1.4 mm / .055 in. This is approximately 15 fins
per inch.
The spraying density was specified, which allowed the determination of the leaving
concentration and outside heat transfer coefficient using the empirical relationships derived
from Figure 3-21 and Figure 3-22. The subscripts ft and st represent finned tube and smooth
tube.
54

5 2
60 011 1 4
o, ft
x . . e
= +
3 5
60 9 9 1 3
o,st
x . e .
[ 3-51 ]

2
e

= + [ 3-52 ]

where x
out,ft
is the exiting concentration for finned tube[%]
is the spraying density [lbm/hr-ft]

The following sets of equations are derived from Figure 3-23. The subscript o represents
outside and g represents generator.

26
160
.
og , ft
h = [ 3-53 ]

23
120
.
og ,st
h = [ 3-54 ]

With h
og
known the generator pressure can then be determined from the empirical
relationship derived from Figure 3-21.
[ 3-55 ]
2
415 970 727
og , ft g , ft g , ft
h P = + P
[ 3-56 ] 731 342
og ,st g ,st
h P =

where P
g
is the equilibrium pressure in the generator [psia].

With the pressure known, the temperature of the entering and exiting solution can be
determined from property relations and thus an average viscosity can also be determined
from property relations. The main objective of determining the average viscosity is to relate
the outside heat transfer coefficient to a film Reynolds number.
55

4
Re
g
fg
fg
= [ 3-57 ]

where
g
is the film flow rate
7
2
g
tube g
m
Le NC
=

[lbm/hr-ft]
Le
tube
is the length of generator tubes [ft].
NC
g
is the number of tube columns.
7
m is the mass flow rate of the weak solution entering the generator[lbm/hr].

The main difference between film flow rate and spraying density is the film flow rate is the
spraying density divided by two, because the flow is equally split between the tubes.

Figure 3-24 shows the outside heat transfer coefficient as a function of the film Reynolds
number.

Figure 3-24. Outside heat transfer coefficient as a function of Reynolds number.
56

Equations 3-58 and 3-59 are the film Reynolds correlations. The power series was used
because it has the best R
2
value of the forms investigated and is the common relationship
used by the industry.

.24
,
145Re
g st fg
ho =
.27
,
199Re
[ 3-58]

g ft fg
ho = [ 3-59]

The film Reynolds correlation is then compared with empirical relationships for multi-
component mixtures and turbulent falling films from Rohsenow et al., (1998). Figure 3-25
illustrates the large difference between the correlations of Wang and Rohsenow. The
Rohsenow correlation for multi-component mixtures was developed for an ethylene glycol /
water mixture, which might explain why there is a difference. The turbulent correlation takes
into account wave properties of the falling film but it is based on a single component
solution.
57

Figure 3-25. Comparison between outside heat transfer coefficient for different correlations.

The Wang heat transfer coefficient is an average value across the entire length of tube. It
should also be noted that this correlation will be used for a tube bank that is larger than 8
columns and 1 row, which means the correlation does not take into the account the effect of
heat transfer as the falling film proceeds down the tube column.

58

CHAPTER 4

MODEL DEVELOPMENT OF HALF EFFECT CYCLE

4.1 Overview

Figure 4-1 is a schematic of the half-effect cycle. The major difference between the single
and half effect cycle is the addition of a second generator and absorber. This addition allows
for an extra operating pressure at the low concentration [LC] absorber and low temperature
[LT] generator.

The principle of the half effect cycle is that it has two lifts. The term lift is used to represent
a concentration difference between the generator and absorber. This concentration difference
is what drives or gives the potential for mass to flow into the absorber. With the single effect
there is only one lift and thus as the hot water temperature is decreased the difference
between the two concentrations decreases. The two lifts enable the half-effect cycle to
operate at lower firing generator temperatures.

59

chilled
water
Evaporator HC Absorber
(liquid spill)
LTHX
Cooling
Water
Cooling
Tower
HT Generator Condenser
LT Generator LC Absorber
HTHX
Heat
Source

Figure 4-1. Schematic of the half-effect cycle.

The disadvantage of the half-effect cycle is that the COP is approximately half that of the
single-effect and thus more thermal energy is required to produce a specified cooling
capacity. Consequently higher rates of heat rejection are required for a specified capacity.
The higher heat rejection necessitates a larger cooling tower, which translates to higher
operating cost and a larger capital investment. The advantage is that the half-effect cycle can
operate with the same capacity using lower firing temperatures and thus the heat source has
the potential of being free.

Figure 4-2 represents a Dhring plot of the half effect cycle with the corresponding
components. The sloped lines represent constant concentration of LiBr in water. The y-axis is
60

the equilibrium temperature of the water vapor being absorbed or boiled off and the x-axis is
the equilibrium temperature of the LiBr-H
2
O solution.
The absorption process on the Dhring plot shows a process at constant refrigerant
temperature and concentration. The process at constant refrigerant temperature occurs in the
absorber and generator. The constant concentration lines describe the processes occurring
when the solution is passing through the heat exchanger.

Figure 4-2. Dhring plot of the half-effect cycle.

The important distinction between the single and half-effect cycle can be seen in Figure 4-3,
for the single-effect there are two paths of constant refrigerant temperature where the half-
effect has three.
61

Figure 4-3. Dhring plot comparison between the half-effect and single-effect cycle.

This difference allows the half-effect cycle to produce a larger overall concentration
difference compared with the single-effect. The greater the separation between the
concentration lines the greater the lift and thus the ability to generate cooling. Since both
plots are at 600 tons, the greater concentration separation allows the half-effect cycle to have
the sum of the UA for all components to be less than the single-effect, which also means less
tube area. The ability for the half-effect cycle to operate using a lower temperature heat
source can be seen in Figure 4-3 by comparing the solution temperature. The HT
gen
solution
temperature is less than the single-effect generator solution temperature because the
concentration is less even though the pressure or refrigerant temperature is slightly greater.
Also the LT
gen
solution temperature is less than the single-effect because the pressure or
62

refrigerant temperature is less even though the concentration is the same. By having a lower
generator solution temperature enables the half-effect to operate with lower inlet hot water
temperatures.
4.2 Inputs

The same inputs are required for each component as for the single-effect model. The main
difference between the half-effect and single-effect model is that an additional mass and salt
balance is required. The following is a complete list of the inputs for the half-effect cycle.
4.2.1 HTGenerator

T30 Temperature of hot water [F]
Vol30 Flow rate of hot water [gpm]
NCG_ht Number of Tubes in column
4.2.2 LTGenerator

Vol
31
Flow rate of hot water [gpm]
NCG_lt Number of Tubes in column

Signal_hs Control variable to determine if temperature entering is leaving HT generator
temperature or if a defined temperature.
0 split flow and 1 series flow
4.2.3 Condenser

Le_tube Length of Tubes [ft]

63

Signal Control variable used to determine if cooling water flow rate and temperature
is equal to absorber flow rate and leaving LC absorber
0 closed valve [specify temperature and flow rate]
1 open valve no input required
4.2.4 Evaporator

NCE Number of Tubes in a column

m50 Mass Flow Rate of Chilled Water [lbm/hr]
T51 Temperature of Delivered Chilled Water [F]

4.2.5 HC Absorber

T40 Temperature of Entering Cooling Water [F]

NCA_hc Number of Tubes in a column

4.2.6 LC Absorber

NCA_lc Number of Tubes in a column

4.2.7 High-Temperature Heat Exchanger

_hthx Effectiveness of High-Temperature Heat Exchanger

64

4.2.8 Low-Temperature Heat Exchanger

_lthx Effectiveness of Low-Temperature Heat Exchanger
4.3 Assumptions

The same assumptions are applied to the half-effect cycle model as for the single-effect
model (section 3.3) with the following differences

The sub-cooling and liquid spill in the single-effect model is not applied to the half-effect
cycle. This gives a small advantage to the half-effect cycle, but since no data is available it
would be unfair to penalize a cycle without knowing if sub-cooling or liquid spill would
occur. If it turns out that the half-effect cycle is not economically viable in the ideal case then
why go into the detail of sub-cooling and liquid spill.

The additional assumption is that the refrigerant flow rate leaving the low temperature
generator is equal to the refrigerant flow rate leaving the high temperature generator. Since
the model is steady state and does not allow for storage of mass, this assumption must be
implemented to ensure mass conservation. Each of the components in the half-effect cycle
are modeled using the same equations as described in section 3.4.
65

66

CHAPTER 5

COOLING TOWER

5.1 Overview

The cooling tower model is needed to predict parasitic fan power for different design and
operating characteristics of the absorption chiller. As the cooling water flow rate is
increased, the absorber has a higher overall heat transfer coefficient [UA
abs
], but it comes at a
cost because a larger cooling tower is needed. The cooling tower model is based on the
analogy approach developed by Braun et al. (1989). A program called UPDATE 3.34,
(Marley Cooling Tower Company, 1998) generated performance curves, which were used to
compute the cooling tower overall heat transfer coefficient [UA
tower
], needed for the analogy
approach method. Figure 5-1 displays a schematic drawing of a cooling tower with the inputs
and outputs used in the model.

67

Water
Air
Tower
T
wi
, VA
w
[gpm]
T
wb
,VA
air
[cfm]
UA
T
wo
P
fan
Water makeup

Figure 5-1. Schematic diagram of a cooling tower.

The inputs to the model
T
wi
Inlet water temperature [F]
VA
w
cooling water flow rate [gpm]
T
wb
Wet bulb temperature [F]
UA
tower
Tower UA [Btu/hr-F]
T
wo
or P
fan
Outlet water temperature or Fan Power [F] or [hp]

The last input enables the model to either hold fan power constant and have the outlet water
temperature vary with wet-bulb temperature or require a constant cooling water temperature
and vary the fan power. The water makeup is an output of the model although it is physically
an input to the cooling tower, due to the water loss into the air stream.

68

5.2 Analogy Approach

The analogy approach is based on the same principles as a sensible heat exchanger cooling
coil. The following description is a condensed version from Braun et al. (1989). The mass
and energy balance on the cooling tower is:

[ 5-1 ]
w,o w,i a,o a,o a,i w,i
m m w m w m =
[ 5-2 ]
, , , , , ,
( ) (
w w o w o w i w i a a o a i
C m T m T m h h = )

where is the water or airflow rate [lbm/hr] m
w is the humidity ratio of the air [lbm
w
/lbm
dry,a
]
C
w
is the specific heat of water [Btu/lbm-F]
h is the enthalpy [Btu/lbm]
T is temperature [F]

w
represents water

a
represents air

i,o
represent in and out

The effective specific heat is the change in enthalpy with temperature along the saturation
line. The effective specific heat can be used to rewrite the energy equation is terms of
enthalpies.

a,sat ,i a,sat ,o
s
w,i w,o
h h
C
T T
[ 5-3 ]

where h
a,sat
is the air saturation ethanlpy at the water inlet and outlet temperature.
C
s
is the effective specific heat

Thus the equivalent capacitance rate is defined as
69

* a s
w,i w
m C
m
m C
=

[ 5-4 ]

The effectiveness is defined similar to a counter-flow heat exchanger except C
*
is replaced
with m
*
.

1
1
1
1
*
*
Ntu( m )
* Ntu( m )
e
m e
[ 5-5 ]

With an effectiveness relationship known the heat transfer is given by

a a,sat ,i a,i
q m ( h h ) = [ 5-6 ]

The outlet air conditions are based on an air passing over a uniform wetted surface at a
specified uniform temperature. The energy equation for this effective enthalpy is

1
a,o a,i
a,eff a,i
Ntu
h h
h h
e
= +
[ 5-7 ]

An outlet air humidity ratio is also determined by a similar approach
[ 5-8 ]
Ntu
a,o s ,eff a,i s ,eff
w w ( w w )e
= +

An energy balance on the air side yields
[ 5-9 ]
a a,o a,i
q m ( h h = )

All of the above equations are based on a general cooling tower design. The Ntu is
determined from a UA, which is based on the characteristics of an actual cooling tower. The
Ntu is defined as

a a
UA
Ntu
m C
=
[ 5-10 ]
70

5.3 Fan Power

It is assumed that the fans obey the cubic power law and thus the power consumed as a
function of air flow rate can be represented by
[ 5-11 ]
3
1
cell
N
fan i fan,nom,i
i
P P
=
=

where
air
air ,nom
VA
VA
=

The nominal air flow and fan power is based on standard conditions of 78
o
F wet bulb, 95/85
o
F inlet and outlet water temperature, and 3 gpm/Ton.

The fan blade mechanical efficiency at part load conditions is given by the function
[Koeppel, 1994]

2
1 3323 15 2 8047 2 756 90133
mech
. E . . .
3
= + +
4
[5-12 ]

Koeppel (1994) suggest a curve fit for a pump motor/drive efficiency in the calculation of the
fan motor efficiency of the form:

2 3
-.00031247+2.1943 -.49874 -1.7454 +1.0504
motor
= [5-13 ]
71

Thus the electrical power consumption is

fan
elec
motor elec
P
P

=

[5-14 ]

5.4 Overall Tower Heat Transfer Coefficient

The UA for a cooling tower was determined from performance curves provided by UPDATE
3.34 (Marley Cooling Tower Company, 1998). The performance curves are fan power [P
fan
]
versus wet bulb temperature for 3 different ranges. The range is defined as the temperature
difference between the water entering and leaving the cooling tower. The three ranges are 8,
10 and 12
o
F. The UA
tower
was determined by inputting the fan power to the model and having
the analogy approach calculate the UA
tower
for that fan power at a given wet bulb
temperature. Since the program only provides graphs, the fan power and wet bulb
temperature values were read from the graph. The graphs provided by UPDATE 3.34 are
presented as guaranteed performance curves.

Figure 5-2 displays UA
tower
as a function of airflow rate for a constant range and capacity.
The airflow rate is used because there is not enough information available on the geometry to
take a Reynolds number approach. The empirical linear relationship [UA
tower
Model]
developed from the performance data is used to determine the UA
tower
for one type of cooling
72

tower. The empirical relationship was developed out of a need to determine the fan power for
part load conditions. The relationship is of the form

a
tower
cells
mVol
UA b
N
= + [ 5-15 ]

where m and b are constants determined from the performance data [Btu-min/hr-F-ft
3
] and
[Btu/hr-F]
N
cells
is the number of cells in the cooling tower
Vol
a
is the air flow rate [cfm]

Figure 5-2. UA
tower
versus air flow rate for 10 Range, T
wo
=85
o
F , T
wb
=78
o
F, VA
w
=2000
[gpm]

73

The same type of liner relationship is used to determine the UA
tower
for 4 other cooling
towers. The different relationships are needed because as the cooling water flow rate is
varied different cooling tower sizes are needed. For example choosing only one cooling
tower and varying the flow rate would greatly under or over design the system. Table 5-1 list
the different types of cooling towers used in the system model.

Table 5-1. Different types of cooling towers available
Model VA
air
[cfm]
per cell
VA
w
[gpm] Cost
Factor
cells FanPower
max

[hp] per cell
NC9202 176400 4000 1.31 2 25
NCB222 190100 4000 1.00 2 40
NCB202 164300 3500 1.09 2 25
NCA202 163700 3000 1.13 2 25
NCC221 235400 2500 1.10 1 40
NC9201 176400 2000 1.31 1 25

The cost factor is the relative increase in price due to a larger cooling tower with less fan
power. The importance of the cost factor is that there are cooling towers available that can
meet the same load with different fan powers, thus the cost factor can be used to see if the
lower operating cost compensates for the increase in heat transfer area which translates to a
higher capital investment.

5.5 Calibration/Prediction

Figure 5-3 and Figure 5-4 demonstrate the ability of the analogy approach to predict fan
power and outlet water temperature as a function of wet bulb temperature. The performance
curve data is read off of graphs generated by UPDATE 3.34.
74

Figure 5-3. Fan Power versus wet bulb
temperature [4000 gpm, NC9202]

Figure 5-4. Water outlet temperature
versus wet bulb temperature [4000 gpm,
NC9202]

The model was calibrated for a temperature range of 10
o
F and is able to predict a 12
o
F
temperature range with the same accuracy. The above graphs demonstrate the linear
relationship of UA
tower
as a function of airflow rate and the analogy approach are valid for
modeling a cooling towers power consumption and outlet water temperature.

75

76

CHAPTER 6

ECONOMICS

6.1 Overview

Economics is the parameter that bridges the gap between a thermodynamic analysis and the
power required to operate the unit. Thermodynamics indicates that the optimal system will
have the highest overall heat transfer coefficient [UA]. However, in reality economics puts a
limit on what the UA can be. For example, it is possible to increase the capacity by
increasing the cooling water flow rate. However, an increase in cooling tower water requires
a larger cooling tower and higher operating cost. The purpose of this section is to set up the
assumptions and cost figures that will be used to optimize the absorption chiller system. Two
different economic analyses are performed using the P
1
, P
2
method (Beckman and Duffie,
1992). The first analysis determines the best design of absorption chiller and cooling tower
system. The second analysis determines the cost of heat needed to make an absorption
chiller competitive with an electric centrifugal chiller
77

6.2 Material / Installation Cost Assumptions

6.2.1 Absorption chiller

The material cost for the absorption chiller is primarily based on the number of tubes, shell
area, and pumps. These three components have the greatest effect on dollars per ton of
cooling capacity for a change in cooling water temperature, or heat source temperature, and
flow rate. For example, a lower heat source temperature will require a larger UA in the
generator to meet a fixed capacity. The purging system, salt solution, control system, piping,
etc., are relatively minor costs and are not effected by a change in the heat exchanger design.
A constant price of $20,000 will account for the above components.

The following cost parameters reflect the figures used to determine the capital cost of an
absorption chiller (US chiller manufacturer, 2000). Table 6-1 lists the cost of each tube in
the various components.

Table 6-1. Price Range for the tube cost of each component (US chiller manufacturer, 2000).
Component Low [$/ft] High [$/ft]
Absorber 0.35 0.55
Evaporator 0.8 1.0
Generator [Smooth] 0.4 0.6
Generator [Finned] 0.8 1.2
Condenser 0.5 0.7

78

Rather than taking into account how the shell geometry will change by increasing the number
of tubes, it is assumed that the shell, tube sheet, tube supports, and solution distribution
network of the generator-condenser are 75%, and those for the evaporator-absorber are 65%
of the total cost of the tubes. A base case cost of tubes is used to calculate this cost, thus
when the finned tubes are used instead of smooth, the cost of the shell does not increase with
the cost of the tubes but rather with the number of tubes. The installation cost for the tubes is
assumed to be $4 per tube. The installation takes into account the drilling and welding for
each tube. The labor and overhead for the construction of the absorption unit is assumed to
be 65% of the total material cost (US chiller manufacturer, 2000). The installation of the
absorption unit itself is 25% of the total capital cost (RS Means Co. Inc., 1998).

6.2.2 Electric Centrifugal Chiller

RS Means Co. Inc., (1998) lists a cost for a water-cooled electric centrifugal chiller at 300
$/ton for 600 tons and an absorption unit cost of 475 $/ton for 600 tons, while the US chiller
manufacturer (2000) quotes the latter price at 200 $/ton. To stay consistent with the
manufacturers data for an absorption units tube and shell cost, the capital cost for the
manufacturers electric chiller will be 150 $/ton. The installation cost is roughly 30% of the
capital cost (RS Means Co. Inc., 1998).

79

6.2.3 Cooling Tower

RS Means Co. Inc., 1998 quotes the capital cost for the cooling tower at 62$/Ton for
galvanized steel and 100 $/Ton for the stainless steel. Within this range, the capital cost of 70
$/Ton, quoted by Fairchild, (2000), is used. Since it is possible to have more than one
cooling tower meet the demand with different fan power consumption, a cost factor
adjustment is used. The cost factor takes into account the larger unit with smaller fan power.
Refer to the chapter 5, Table 5-1 for the cost factors that are used for different cooling tower
models. The installation cost for a cooling tower is 50% of the capital cost (Fairchild, 2000).
The installation is based on piping, electrical connection, and foundation.

6.3 Operating Cost Assumptions

The operating cost for the absorption system consists of the cooling tower fan and pumps.
The pumps include cooling water, heat source, evaporator, and internal brine solution leaving
the absorber. It is important to take into account the pumps because by increasing the cooling
or heat source flow rate, the absorption chiller capital cost will decrease while the operating
cost will increase.

In analyzing the best absorption system, the heat source is assumed to be free, thus no cost
for the heat is included. The model does not take into account any adverse effects on the
equipment from where the heat source comes from. For example, if the cooling jacket of a
diesel generator is used for the waste heat, it may change the performance of the diesel
80

generator necessitating more fuel to meet the electrical demand. The model does not take into
account maintenance cost for the absorption or electric centrifugal chiller.

6.4 Life Cycle Cost Approach

The economic parameters are P
1
, P
2
, time period of investment, and electrical cost. Table 6-2
lists the numerical range for each one of these parameters. The following section
demonstrates how P
1
and P
2
are determined.
Table 6-2. Range of numerical parameters.
P
1
P
2
Time Period Electrical Cost
Time Period /2 0.8-1.2 10-20 years .05-.2 $/kW

A life cycle cost analysis is performed on the absorption and electric centrifugal chillers. The
life cycle cost combines present worth factors into two variables named P
1
and P
2
. P
1
takes
into account fuel inflation rate and discount rate. P
2
takes into account taxes, mortgage rates,
maintenance, insurance, etc, for a fraction of the initial investment. The following life cycle
cost equations are used for the analysis to determine the cost of heat. The subscripts used in
the nomenclature are

ec is electric centrifugal chiller
ab is absorption chiller
sp

is solution pump
ct is cooling tower
fan is the cooling tower fan

81

[ 6-1 ]
1 2
1 3 1 5
ec ec ec, fan e ec ec,ct
LCC P( P P )C P ( . C . C ) = + + +

[ 6-2 ]
ab 1 1 2
= 1
ab, fan sp evap e hs h ab,ct ab
LCC P( P P P )C PQ C P ( . C . C ) + + + + +

5 1 2

where P
1
is the ratio of life cycle fuel cost savings to first year fuel cost [~ N
e
/2]
P
2
is the life cycle expenditures because of the capital investment. [~ 0.8-1.2]
P

is pumping or fan power [kW]
C
e
is the cost of electricity [$/kW]
C
h
is the cost of heat [$/Therm]
C is capital investment [$]
Q
hs
is the heat source input [Therms]

The definition of P
1
is (Duffie and Beckman, 1991)

1
(1 ) ( , , )
e f
P Ct PWF N i = d [ 6-3 ]
where t is the effective income tax rate.
PWF is a present worth factor
N
e
is the period of economic analysis [year]
i
f
is the fuel inflation rate.
d is the discount rate.

Businesses are able to deduct fuel cost and the effective income tax rate is about 50%. If the
effective interest rate equals the fuel inflation rate then P
1
is approximately equal to N
e
/2. If
the fuel inflation rate is less than the discount rate then P
1
is less than N
e
/2.

The definition of P
2
is (Duffie and Beckman, 1991)
82

min,
2 min,
min,
'
min
( 0, )
1
(1 ) (1 ) ( , )
( , 0, ) ( , 0, )
( 0, )
( , 0, )
(1 ) ( , , ) (1 ) ( , , )
( , 0, ) (1 )
(1 )
e
L L
L
s e e
v
N
D
PWF N d
P D D t D PWF N m d m
PWF N d PWF N m
PWF N d
PWF N m
M t PWF N i d tV t PWF N i d
R Ct
PWF N d Ct
N d
| |
= +
|
\ .
(
+
(
+ +

+
[ 6-4 ]
where D is the ratio of down payment
N
L
is the term of the loan [years]
N
min
is the year of mortgage payments min(N
L
or N
e
) [years]
m is the annual mortgage interest rate
M
s
is the ratio of first year cost (maintenance, insurance) to initial investment
t is the property tax based on assessed value
V is the ratio of assessed valuation of the system in first year to the initial investment
N
D
is the depreciation lifetime [years]
N
min
is the years which depreciation contributes min(N
D
or N
e
).
R
v
is the ratio of the resale value.

If D is set to one and there is no maintenance cost, resale value, assessed valuation, or
depreciation then P
2
is equal to 1. A value less than one for P
2
means there is depreciation,
or a resale value or part of the initial investment is made through a loan. A value greater than
one for P
2
means there are maintenance costs or property taxes.

If the heat source is free then a maximum capital investment for heat recovery is determined
by solving

[ 6-5 ]
1 2
0
hs h hr
PQ C PC =

where C
HR
is the capital investment necessary for heat recovery [$]

83

The result of setting equations 6-1 and 6-2 equal to each other will determine the bread even
cost of heat. The result from equations 6-1 and 6-2 is used in conjunction with equation 6-5,
to calculate an equivalent heat recovery investment based on the breakeven cost of fuel. For
example, the capital investment for heat recovery could be a heat exchanger for the exhaust
of a diesel generator, gas turbine, or industrial stack gases.
84

CHAPTER 7

SINGLE-EFFECT ABSORPTION CHILLER SYSTEM OPERATION

7.1 Overview

The single-effect absorption chiller system encompasses the cooling tower, heat exchanger,
pumps for cooling water and heat source, and the absorption chiller. Figure 7-1 is a
schematic diagram of the absorption chiller system.

85

chilled
water
Generator
Condenser
Evaporator
Absorber
(liquid spill)
LTHX
Cooling Tower
Heat Source / EHX
HSP
CTP
SP

Figure 7-1. Single-Effect Absorption Chiller System.

7.2 Optimization of the absorption chiller system for a given heat source
temperature

The first step in designing an absorption chiller using low temperature waste heat is to
determine which component will have the largest change in capacity for an equivalent
change in UA. Table 7-1 lists the base case for the single-effect absorption chiller

Table 7-1. Base case component size and internal solution flow rate [200 F @ 2000 gpm]
Absorber Condenser Evaporator Generator
Tubes 514 128 336 262
UA [Btu/hr-ft
2
-F] 535304 877538 1.4e6 309516
Internal solution
flow rate
230 [gpm]
86

The design condition of 600 tons is based on the fact that the component sizes in Table 7-1
can obtain this capacity with 27 psia steam. The capacity for the base case at 200
o
F hot water
is 418 tons, thus the component sizes must be changed to meet the design conditions.

Figure 7-2. Base case Capacity and COP versus entering hot water temperature.

Figure 7-2 displays the need for design changes in the base case absorption unit because as
the entering hot water temperature drops to 225
o
F, the capacity falls below design
conditions. The COP in is independent of the entering hot water temperature.

Figure 7-3 displays a plot of the factor change in each component and the corresponding
change in capacity for the base case. Holding 3 out of the 4 component sizes constant while
allowing capacity to change was the method used to produce Figure 7-3. The generator has
87

the steepest slope of the four at 1, therefore any investment should first be put into the
generator.

Figure 7-3. The effect of capital dollars per ton for a factor change in UA.

It is possible that an optimum could be found by changing the generator and absorber,
however, for this analysis only one component at a time is changed. This method allows one
to see the effect of each component separately rather than try to do all combinations
simultaneously. Another result of Figure 7-3 is that the evaporator and condenser show the
least increase in capacity for an increase in the UA.

The optimization of the absorption chiller system includes minimizing capital and operating
cost for the cooling tower, pumps, and absorption chiller. The optimization uses equation 6-
88

2, with the exception of C
h
being set to zero. The analysis involves finding the minimum
value of LCC
ab
. The flow rates are used as the decision variables because they have the
largest effect on capital and operating cost. Figure 7-4, Figure 7-5, and Figure 7-6 show
contour plots of the volume flow rate of heat source and cooling water with the life cycle cost
[LCC
ab
] as the contour variable for 200
o
F hot water, for various values of P
1
. The capacity
is held constant and the number of tubes in the generator changes to find the optimum
generator size based on cooling water and heat source flow rate. The condenser, evaporator,
and absorber are sized for the base case [Table 7-1]. The type of cooling tower also changes
with regard to flow rate. For example, at 4000 gpm, a larger cooling tower is used compared
to 3000 gpm.

Figure 7-4. Contour plot of PV for 200 F entering hot water, P
1
=1.0
89

1
=0.8

1
=1.2
90

Figure 7-4, Figure 7-5, and Figure 7-6 demonstrate that changing P
1
only changes the LCC
and not the position of the optimum solution. The optimum solution is around 3100 gpm for
the cooling water and 2600 gpm for the hot water flow rate.

The above graphs are based on an electrical cost of 0.05 $/kWh, thus Figure 7-7 and Figure
7-8 display the effect of an increase in electrical cost to 0.1 and 0.2 $/kWh.

Figure 7-7. Contour plot of PV for 200 F entering hot water, 0.1 $/kWh electrical cost
91

Figure 7-8. Contour plot of PV for 200 F entering hot water, 0.2 $/kWh electrical cost

One interesting result is that a local minimum occurs at a higher cooling water flow rate. The
second minimum occurs because the cooling tower operating cost is the same for 3600 gpm
and 3100 gpm. The difference is the capital cost of the cooling tower; Figure 7-7 and Figure
7-8 reveal the second minimum at 3600 gpm. Figure 7-4, Figure 7-5, and Figure 7-6 also
show a second minimum, but the resolution is not small enough to show the minimum in the
graphs. Figure 7-6 discloses a small sign of the second minimum due to the indention at 3600
gpm.
92

2
=5 or 10 year life cycle.

Figure 7-9 shows the contour plot when the life cycle is 10 years instead of 20, as in Figure
7-4. The optimum solution is near 3100 gpm for the cooling water and 2900 for the heat
source flow rates.

With a change in P
1
, P
2
, the cost of electricity, and the life cycle time, the optimum solution
flow rates for the absorption chiller system remained the same. The only change is in the
value of the LCC
ab
. The importance of the above parameters will be discussed when
comparing this system to an electric centrifugal chiller.
93

Figure 7-10. Contour plot of PV for 230
o
F entering hot water.

Figure 7-10 shows the effect on the contour plot by changing the inlet hot water temperature.
The optimum solution for a higher inlet hot water temperature moves the flow rate from the
right corner to the left. This result was expected because a higher temperature heat source
requires less flow rate for the same capacity.

From Figure 7-4 to Figure 7-10 the optimum value for the cooling water flow rate is
independent of the heat source temperature. The range of 4000 to 3000 gpm was chosen
based on numerous runs to investigate possible solutions. Below 3000 gpm the capital cost is
extremely high and anything above 4000 gpm is too large of a flow rate. Likewise, for a
94

lower heat source temperature a larger flow is advantageous because it reduces capital cost
more than the extra operating cost for the larger pump.

7.3 Effect of heat source temperatures

Firing the absorption chiller at lower temperatures is possible but it comes at a cost. The
lower temperatures require a larger unit, increasing the capital cost. Figure 7-11 and Figure
7-12 examines the effect of UA and capital dollars per ton.

Figure 7-11. Capital dollars per ton as a function of entering hot water.

95

Figure 7-12. UA generator as a function of entering hot water.

The UA and dollars per ton reveal the same trends, however, it is important to look at both
the cost and the UA. The dollars per ton is calculated by using empirical heat transfer
coefficients and pricing assumptions based on material and labor costs [Refer to chapter 3
and chapter 6]. The UA is calculated by energy, mass balances, and property data, requiring
fewer assumptions and making it less uncertain.

Figure 7-12 shows that at temperatures below 205
o
F, the chiller design starts to change
rapidly to accommodate lower firing temperatures. At around 210
o
F the UA
generator
nearly
doubles in size compared to the base case in Table 7-1. As the entering temperature is
lowered the effect of hot water flow rate becomes more important in the design.

96

The capital dollars per ton is based on material, labor and overhead cost for the absorption
chiller presented in 6.2. Figure 7-11 shows the same trends as Figure 7-12, but this can be
used as a tool to compare with other first cost cycles and a standard parameter used by the
refrigeration industry. At 200
o
F, a 33% increase from 1500 to 2000 gpm in heat source flow
rate can reduce the capital cost by approximately 20% from 260 to 220 $/Ton, assuming one
can acquire a heat source with a flow rate of 2000 gpm for a reasonable investment.

7.4 Sensitivity Analysis on UA of each component.

In order to meet the 600 tons at 200
o
F entering hot water the base case designed must be
changed. Table 7-2 lists the absorption chiller design to meet the 600 tons of cooling.

Table 7-2. Component size and internal solution flow rate to meet 600 Tons [200
o
F @ 2000
gpm]
Tubes 514 128 336 1730
UA [Btu/hr-ft
2
F] 522470 853445 1.4e6 1.35E6
Internal solution
flow rate
230 [gpm]

Table 7-2 shows about a 650% increase in the number of generator tubes, compared to Table
7-1 to deliver 600 tons at 200
o
F entering hot water. In Table 7-2 only the generator size was
changed, thus a sensitivity analysis on each component UA is needed to obtain a more
effective design. The capacity is held constant and the number of generator tubes is changing
as the absorber, evaporator, and condenser UA is increased and decreased.

97

Figure 7-13. The effect of capital dollars per ton for a factor change in UA.

Figure 7-13 demonstrates that increasing the size of the absorber and correspondingly
decreasing the size of the generator will have the largest effect on cost while maintaining a
fixed capacity. Any investment made into improving the design of the low temperature hot
water absorption chiller should be put into the absorber first. Table 7-3 shows that by
increasing the number of tubes in the condenser, it will reduce the greatest amount of
generator tubes. Since there are fewer tubes in the condenser, a 2% increase will result in an
increase of 3 tubes for the condenser where as 10 for the absorber. The disadvantage is that
the condenser tubes are more expensive than absorber tubes, making it more economical to
increase the absorber tubes [see Table 6-1 for tube cost information].

98

Table 7-3. The effect of increasing the tubes in the absorber, condenser, and evaporator with
the corresponding decrease in generator tubes.
%Change in
UA
Tube
a
Tube
g
Tube
c
Tube
g
Tube
e
Tube
g

2% +10 -150 +3 -74 +6 -40
4% +20 -270 +5 -140 +14 -75
6% +30 -370 +8 -200 +20 -108
8% +40 -454 +10 -250 +27 -138
10% +50 -526 +13 -300 +34 -165

Table 7-3 indicates that if 10 tubes are added to the absorber, 150 tubes can be taken out of
the generator.

The low temperature heat exchanger was assumed to have 0.76 effectiveness for the above
analysis. The effect on capacity and cost for increasing the effectiveness was investigated.
The effectiveness of the brazed plate heat exchanger is increased by adding more brass plate.
Figure 7-14 and Figure 7-15 compares the effect on capacity and unit cost for a change in the
effectiveness of the solution heat exchanger.

99

Figure 7-14. Capacity effect due to changing LTHX effectiveness.

Figure 7-15. Capital dollars per ton due to changing LTHX effectiveness

Using an effectiveness of 0.8 increases the capacity by around 15 tons and at no extra cost
since the generator can be made slightly smaller.

100

Another approach to decrease the number of generator tubes is to use finned tubes in place of
smooth tubes. Figure 7-16 and Figure 7-17 displays the decrease in the number of tubes and
the cost associated with making such a change.

Figure 7-16. Comparison between finned and smooth tubes.

101

Figure 7-17. Cost Comparison between finned and smooth tubes.

Figure 7-16 shows that using finned tubes decreases the number of generator tubes. Based on
the assumption that finned tubes cost twice as much as smooth tubes, the reduction in
generator tubes is enough to compensate for the higher priced tubes. Figure 7-17 shows that
the overall design price is much lower for the finned tubes, making it economically and
thermodynamically advantageous to use finned tubes in a generator fired with low
temperature hot water.

With falling film generators there is the potential for the exiting brine solution leaving the
generator to be superheated. Figure 7-18 and Figure 7-19 displays a comparison between no
superheat and a 3
o
F superheat for smooth and finned tubes. Superheat is the condition where
the exiting brine solution leaving the generator is at a higher temperature then the equilibrium
102

temperature for that concentration and pressure. Due to the loss of performance caused by
superheat in the generator, there is an increase in the cost.

Figure 7-18. Capacity effect between finned and smooth tube with and without superheat.

Figure 7-19. Capital dollars effect between finned and smooth tube with and without superheat.

103

The conclusion of Figure 7-18 and Figure 7-19 is that the finned tube generator is superior to
the smooth tube with or without superheat.
Table 7-4 lists the final design parameters for a single-effect chiller using low temperature
hot water. An 1133-tube count in the generator is still not a feasible design for the generator.
The major changes in the final design are the increase in absorber tubes from 514 to 554 and
the slight increase in effectiveness of the LTXH from 0.76 to 0.8.

Table 7-4. Lists the final design for the absorption chiller using low temperature waste heat
[600 Tons, 200
o
F @ 2000 gpm, no superheat 0.8 LTHX effectiveness].
Tubes 554 128 336 300 Finned
1133 Smooth
UA [Btu/hr-F] 558751 852738 1.35e6 993439
Internal solution
flow rate
230 [gpm]

Figure 7-20 compares the system and absorption chiller cost as a function of entering hot
water temperature based on the final design. The cooling tower cost is based on 70 $/ton and
thus an uncertainty of 20 $/ton is examined in the system curve. The 20 $/ton is used
because it encompasses the cost range of a lower tower cost for galvanized steel and an upper
cost for stainless steel.

104

Figure 7-20. Capital dollars per ton comparison between single-effect absorption chiller and
system.

At 200
o
F and 2000 gpm the absorption chiller and system cost is
222 $/ton 411 $/ton 54.2 $/ton [smoothed tubes]
185 $/ton 374.2 $/ton 54.2 $/ton [finned tubes]

The plus and minus is based on the cooling tower cost range. The important result is that the
cost of the system has approximately doubled due to the cooling tower.

105

7.5 Sensitivity analysis on tube cost and generator outside heat transfer
coefficient.

Since the dollars per ton is such an important parameter used by the industry it is important
to perform a sensitivity analysis. Using the range of tube cost [0.10 $/ft] from the economics
chapter, Figure 7-21 demonstrates the sensitivity of dollars per ton.

Figure 7-21. Uncertainty analysis on the cost of tubes.

Figure 7-21 shows the propagation of uncertainty on the capital dollars per ton associated
with a 0.10 $/ft uncertainty in the price of all four tube components [generator, absorber,
condenser, and evaporator] [Refer to Table 6-1 for tube cost]. The uncertainty increases as
the entering hot water temperature decreases, which means there is a higher uncertainty in
106

predicting the cost of an absorption chiller at lower firing temperature. The reason for this is
that at lower temperatures there is a large increase in the number of generator tubes, thus the
cost of tubes is a greater percentage of the capital cost.

At 200 F the uncertainty range on capital dollars per ton is
Finned $184.6 $3.6
Smoothed $221.5 $6.4

Holding the capacity constant and changing the outside heat transfer coefficient requires a
change in the number of tubes in the generator.

Figure 7-22. Uncertainty analysis on the generator outside heat transfer coefficient.

107

Figure 7-22 shows the propagation of uncertainty on the capital dollars per ton associated
with a 20 % uncertainty in the generator outside heat transfer coefficient. Similar to Figure
7-21 the uncertainty increases as the hot water temperature decreases.

At 200 F the uncertainty range on capital dollars per ton is
Finned $184.6 $4.2
Smoothed $221.5 $12.6

The uncertainty in the outside generator heat transfer coefficient has a larger effect than the
cost of the tubes, especially at lower firing temperatures. With the uncertainty in cost of
tubes and generator outside heat transfer coefficient, Figure 7-21 and Figure 7-22 show that
the finned tubes are still a better design because the confidence intervals do not coincide.

7.6 Life cycle savings comparison between single-effect absorption and vapor
compression chillers for different heat source temperatures

A life cycle savings comparison is used to compare absorption and electric centrifugal
chillers. The break-even cost of heat is determined by setting equations 6-1 and 6-2 equal to
each other. Refer to section 6.2 and 6.3 for material and operating cost details of the
absorption and electric centrifugal chiller. The result reveals the price at which the cost of
heat must be in order for absorption to have an economic advantage over the electric
centrifugal chiller. The final design absorption chiller is used [Table 7-4 finned tube], which
means each component is held constant and the capacity changes with heat source
temperature.
108

Figure 7-23. Break-even cost of supply heat to the generator or heat equipment.

Figure 7-23 demonstrates the break-even cost of supply heat for the generator to operate an
absorption chiller with the same life cycle cost as an electric centrifugal chiller. For
example, if the supply heat can be supplied at a cost less than or equal to approximately 0.28
$/Therm when electricity is 0.1 $/kWh, then absorption is the preferred alternative. The right
hand axis shows when the supply heat is free from an industrial process or co-generation
facility. The right axis determines the maximum break-even cost associated with using waste
heat [refer to equation 6-5 in section 6.4]. For example, at 200
o
F and 0.1 $/kWh, any
investment made in recovering the waste heat that cost less than $2.0 million, is a viable
alternative to the electric centrifugal chiller [based on a 20 year life cycle operating 4380
109

hours a year]. Another interesting conclusion of Figure 7-23 is that the break-even cost of
heat is independent of entering hot water temperature from 190
o
F and up.

Figure 7-24. Capacity and COP as a function of heat source temperature

The main reason for this can be seen in Figure 7-24, where even though capacity is falling
the COP remains constant.

The life cycle savings analysis is based on two important parameters P
1
and P
2
,

thus a
sensitivity analysis is performed due to an uncertainty in P
1
and P
2
. The uncertainty of P
1
is
20% of n where n is life cycle years. The uncertainty in P
2
is 20% with a range of 0.8-
1.2.

110

Figure 7-25. Uncertainty analysis in P
1

Figure 7-26. Uncertainty analysis in P
2

111

Figure 7-25 and Figure 7-26 shows that even with large uncertainties in P
1
and P
2
there is
little effect on the solution. The above graphs are based on a full year of operation, but two
interesting results occur when the absorption unit is only used for half a year. Figure 7-27
compares the uncertainty in P
2
and the cost of heat for 8760 and 4380 hours of operations.

Figure 7-27. Cost of Heat Source uncertainty in P
2
for full and half a year of operation.

The scale is expanded in Figure 7-27 so that the error bars can be compared with Figure 7-26.
The first result of Figure 7-27 is that the break-even cost of heat is greater for a full year of
operation. This increase is due to the operating cost of the absorption chiller being less than
the electric centrifugal chiller. To make use of this benefit, it is important for the absorption
chiller to run as much as possible. The second result is that at 4380 hours per year of
operation, the error in the cost of heat due to an uncertainty in P
2
is larger. One reason for
112

this is that at 4380 hours, the operating cost is lower causing the capital cost to have a larger
effect on the life cycle cost.

If P
1
is less than 1 and P
2
is greater than n/2 then the break-even cost of heat increases which
means a low interest rate or an increase in the economic time period.

Table 7-5. Compares the capital and operating cost between an absorption and electric
centrifugal chiller for a 20 year life cycle [~600 Tons, 0.05$/kWh, 200
o
F Hot Water @ 2000
[gpm], 4380 Hours]
Absorption/ Tower
[finned]
~0.14 $/Therm
Electric Centrifugal /
Tower
Capital Cost
Chiller $136,048 $117,000
Tower $170,591 $81,046
$306,639 $198,046
Operating Cost
Pumps / compressor $11,430 $783,240
Heat Source $642,100
Tower $72,150 $51,030
$725,680 $834,270
Total $1,032,319 $1,032,316

Table 7-5 compares the capital and operating cost for the absorption chiller and the electric
centrifugal chiller. The values in Table 7-5 correspond to one point on Figure 7-23. The total
values in Table 7-5 are the same because the LCC costs are set equal to each other to
determine the break-even cost of heat. The absorption unit is slightly more expansive than
the electric centrifugal chiller but it is the cooling tower for the absorption unit that is about
twice as much compared to the electric centrifugal chiller. The operating cost is not
significantly lower as one would initially guess, for an absorption unit. The main operating
113

cost for the absorption chiller is the cost of heat. Therefore using absorption with co-
generation is the only feasible situation where absorption would be more economically
attractive than current centrifugal chillers, unless there are extremely high electrical costs.

7.7 Summary

Figure 7-28 compares the low temperature hot water design with the base case design. The
major conclusion is an increase in capacity for a lower heat source temperature. The cost of
this new design at 600 tons is 15 $/ton more.
The base case at 600 tons 227
o
F @ 2000 gpm is 170 $/ton
The low temperature design at 600 tons 200
o
F is 185 $/ton

Figure 7-28. Capacity comparison between low temperature hot water design and base case
design.
114

The following is a list of important summaries in Chapter 7

The base case absorption unit starts to fall below design conditions [600 tons] with a hot
water temperature source of 225
o
F @ 2000 [gpm].

Holding capacity constant [600 tons], at 205
o
F the base case chiller size starts to change
rapidly to accommodate the lower firing temperatures.

The generator for the base case absorption unit has the largest increase in capacity cost
for a unit change in UA using a low temperature heat source. Therefore any investment
should first be put into the generator.

The low temperature absorption unit was designed by increasing the number of tubes in
the generator [262 300], using finned tubes instead of smooth tubes, increasing the
number of tubes in the absorber [514 554] and increasing the low temperature heat
exchanger effectiveness [0.76 0.8.].

The optimum value for the cooling water flow rate is independent of the heat source
temperature. The cooling water flow rate is around 3100 gpm or 5gpm/ton at 600 tons.

Based on the assumption that finned tubes cost twice as much as smooth tubes, the
reduction in generator tubes using finned tubes is enough to compensate for the higher
price.

The cost of the absorption chiller system has approximately doubled due to the cooling
tower.

If the heat source can be supplied at a cost less than or equal to approximately 0.28
$/Therm when electricity is 0.1 $/kWh then absorption is the preferred alternative [based
on 600 tons capacity with 190
o
F and up hot water at 2000 gpm].

The break-even cost of heat source is independent of entering hot water temperature from
190
o
F and up.

The main operating cost for the absorption chiller is the cost of heat. Thus, using
absorption with co-generation is the only feasible situation where absorption would be
more economically competitive than current centrifugal chillers.

115

116

CHAPTER 8

HALF-EFFECT ABSORPTION CHILLER OPERATION

8.1 Overview

The purpose of Chapter 8 is to determine an optimum design for the half-effect cycle,
evaluate the performance characteristics of the optimum design, compare this cycle with an
electric centrifugal chiller, and examine different control strategies. Refer to Chapter 4
Figure 4-1 for a schematic diagram of the half-effect cycle and section 4.1 for a review of
basic principals of the half-effect cycle.

8.2 Designing the Half-Effect Cycle

The design approach is to minimize cost based on decision variables of the system. The
decision variables for the half-effect cycle are the number of tubes in each component and
the two solution flow rates leaving the absorber.

117

The first step in designing the half-effect cycle is to size the components comparable to the
single-effect unit to obtain a numerical solution. The condenser and evaporator are initially
set to the same size as the base case unit in Table 7-1. The high and low concentration
absorber sum is set equal to the absorber in the single effect unit [Table 7-1]. The single
effect absorber has about 500 tubes, therefore the high and low concentration absorber will
each have 250 tubes. The solution flow rates leaving the absorber are set to 230 gpm [Table
7-1]. The sizes of the generators are set equal to each other and determined by setting the
capacity to 600 tons.

The half-effect cycle was designed starting with the single-effect unit until convergence on a
numerical solution was reached. The convergence was not a problem of numerical
difficulties but of physical limitations of the system not having enough heat exchanger area
to meet the 600 tons. The procedure followed by setting the solution flow rate leaving the
high concentration [HC] absorber to 200,000 lb
m
/hr and the number of tubes in the absorber
were changed until a solution was reached. The same procedure was implemented for the
low concentration [LC] absorber, by setting the low concentration solution flow rate to
140,000 lb
m
/hr and adjusting the number of tubes. Table 8-1 lists the flow rates and number
of tubes for each component from this analysis.
Table 8-1. Initial Half-Effect Design
HC
Absorber
LC
Absorber
HT
Generator
LT
Generator
Condenser Evaporator
Tubes 300 200 306 306 128 336
Mass
Flow Rate
220 [gpm] 180[gpm]

118

The leaving solution flow rates from the HC and LC have very little effect on the capital
dollars per ton because of the film Reynolds number. The film Reynolds number is just
beyond the Nusselt number limit, thus a change in film Reynolds number does not affect the
Nusselt number. [Refer to Figure 3-3 and section 3.4.3]. Therefore, increasing or decreasing
the solution flow rate over a small range has little effect on the outside heat transfer ability of
the two absorbers.

Figure 8-1 and Figure 8-2 were used to determine the sensitivity of each variable on the
entire system. Holding each variable constant while varying the other produced the
following graphs. For this analysis, the capacity is held constant at 600 tons and the number
of tubes in the high temperature [HT] and low temperature [LT] generator are set equal. This
procedure enables the generator size to vary due to the following different conditions. Figure
8-1 and Figure 8-2 displays the effect on dollars / ton by changing the HC and LC absorber
tubes. There is a small effect on capital dollars per ton for an increase in the number of
absorber tubes. The most likely reason is that since the increase in absorber tubes is small
and as the absorber tubes increase, the generator tubes decrease.

119

Figure 8-1. HC Absorber tubes effect on dollars per ton.

Figure 8-2. LC Absorber tubes effect on dollars per ton.

Because both the mass flow rates have a small effect on the system, and the tube counts of
the evaporator and condenser remain constant there are four variables left that need to be
optimized. The four variables are LT
tubes,
HT
tubes,
LC
tubes,
and HC
tubes
. Using a ratio to
represent the portion of tubes in the HC and LC absorber and HT and LT generator provides
120

limits that are between 0 and 1. For example, the ratio of LC to HC absorber tubes is
optimized instead of the number of tubes. The 2 ratios are f
abs
, f
gen
.

( )
tubes
abs
tubes tubes
LC
f
HC LC
=
+
,
( )
tubes
gen
tubes tubes
LT
f
HT LT
=
+

Specifying capacity allows for HT
tubes
or LT
tubes
not be specified. A value of 350 tubes for
the HC
tubes
absorber will be used for the initial optimization, thus the two remaining variables
to be optimized are f
abs
and f
gen

The objective function is

[ 8-1 ] min($/ ) ( , )
abs gen
Ton f f f =

Where the decision variables are constrained by

0. 1.
abs
f < <
0. 1
gen
f < <

The variables can not equal the constraints because that would result in one of the
components to have zero tubes, which is physically impossible.

The first attempt to solve the optimization problem was to use EES to minimize the dollars
per ton based on f
abs
and f
gen
. The problem with using EES is that within the constrained
variables, the program would take a step that did not have a solution, thus stopping the
program. In order to circumvent this problem, EES would have to know that a solution does
not exist and return to the previous step, solve, and then take a smaller step or change
121

direction. Since the optimization only contains two variables, a contour plot can be used to
find the minimum. The advantage of the contour plot is that the user can control the
variation of f
abs
and f
gen
. Figure 8-3 is a contour plot of capital dollars per ton for different
variations of f
abs
and f
gen
.

Figure 8-3. Contour plot of capital dollar per tons for series flow.

The optimum fractions from Figure 8-3 are
f
abs
= .41
f
gen
= .57

122

Table 8-2 lists the optimum solution for the half-effect cycle using 200
o
F hot water at 2000
gpm in series flow.
Table 8-2. Optimum Half-Effect Design [600 Tons, 200
o
F @ 2000 gpm]
HC
Absorber
LC
Absorber
HT
Generator
LT
Generator
Condenser Evaporator
Tubes 350 243 192 255 128 336
UA
[Btu/hr-F]
375285 327870 246966 283976 882375 1.354e6
Internal
Solution
Flow Rate
240
[gpm]
210
[gpm]

Since the design was based on series flow through the two generators, the same analysis was
performed for parallel flow through the generators.

Figure 8-4. Contour plot of capital dollar per tons for parallel flow.

123

Figure 8-4 is a contour plot of capital dollars per ton for different fractions. The main result
of Figure 8-4 is that the same optimum solution exists regardless of flow arrangement with
the two generators.
Figure 8-5 examines the effect on the optimum fraction by lowering the heat source
temperature to 185
o
F. At lower hot water temperatures the f
abs
is greater than 0.5, which
means the LC absorber is now larger than the HC absorber. On the contrary, the f
gen
did not
change.

Figure 8-5. Contour plot of capital dollars per ton for series flow [185
o
F heat source].

The next step in the design is to determine the sensitivity of the design sizes. For example, if
the condenser size is increased will this decrease the capital cost. Figure 8-6 displays a graph
124

of the sensitivity of each component with regard to capital cost. The capacity is held
constant and a factor is used to increase or decrease one component at a time, which also
effects the size of the generators based on the solved ratio of 0.57.

Figure 8-6. Effect on capital cost due to changing the size of each component.

Figure 8-6 demonstrates that the half-effect cycle is properly designed since there are small
changes in capital cost for a change in UA of each component. Decreasing the size of the HC
absorber below 90% resulted in a design that did not have enough UA to meet the 600 ton
capacity. One possible reason for this result is that as the HC absorber decreases in size, the
leaving concentration increases. Small changes in concentration result in large changes in
solution flow rate, however since solution flow rate is constant, no solution is possible.
125

Figure 8-6 also shows that the evaporator is over designed because the same capacity can be
met with a lower UA for the evaporator.

8.3 Effect of Heat Source Temperature

Figure 8-7 and Figure 8-8 describes the effect on the generator UA for varying hot water
temperatures. The capacity is held constant and the two generators change in size with the
varying hot water temperature. The optimum half-effect fractions are still used but the
number of tubes in each generator is allowed to vary. The UA for the LT generator is
slightly higher than that of the HT, which is a result of the LT generator having more tubes
than the HT generator. Figure 8-9 shows the change in capital dollars per ton for varying
entering hot water temperatures.

Figure 8-7. HT generator UA as a function of entering hot water temperature.
126

Figure 8-8. LT generator UA as a function of entering hot water temperature.

Figure 8-9. Capital dollars per ton for varying entering hot water temperature.

127

As the hot water flow rate increases, the required capital investment decreases. This is the
same conclusion as the single-effect cycle. Figure 8-10 compares the capital cost between an
absorption chiller and the absorption chiller and cooling tower.

Figure 8-10. Capital dollars comparison between half-effect absorption chiller and system

At 200
o
F and 2000 gpm the half-effect absorption chiller and system cost is
275 $/ton 550 $/ton 78 $/ton
The plus and minus is based on the cooling tower cost range. The important result is that the
cost of the system has approximately doubled due to the cooling tower.

128

8.4 Dollars per ton effect due to split or continuous heat source flow rate into
the generators

The above analysis has assumed that the hot water source passes through the HT generator
followed by the LT generator. The following figures were drawn to determine if it is better to
use series flow or parallel flow with half the flow rate. For example, is it more economical to
pass 2000 gpm of hot water first through the HT and then the LT, or pass 1000 gpm each to
both generators? Figure 8-11 and Figure 8-12 compares the two different types of flow
arrangements. The higher the hot water flow rate the lower the entering hot water
temperature can go before there is a difference between series and parallel flow arrangement.
At approximately 190
o
F is it more economical to use series flow rather than parallel flow.
One must keep in mind that the half-effect cycle was designed for 200
o
F water series flow
and Figure 8-3 and Figure 8-4 shows no difference in the optimum fraction. In order for
parallel flow to be as economical as series flow it is best to redesign the half-effect chiller for
a lower-temperature parallel flow arrangement.

Figure 8-11. Effect on capital cost for different flow arrangement 2000 [gpm].
129

Figure 8-12. Effect on capital cost for different flow arrangement 1500 [gpm].

8.5 Life cycle savings comparison between half-effect absorption and vapor
compression chillers

Figure 8-13 displays the break-even cost of the heat source from setting the life cycle savings
equations for a half-effect absorption chiller and an electric centrifugal chiller [equation 6-1
and 6-2] equal to each other. The economic analysis for the comparison between the half-
effect and electric centrifugal chiller is based on the same assumptions in Chapter 6. The
only difference is that the constant price of $20,000, which takes into account the brine
solution, piping, purging system, control system, etc. is doubled due to the extra components
of the half-effect cycle. Similar to the single effect cycle section 7.6, any cost of heat below
the line signifies that the half-effect cycle is economically favored over an electrical
centrifugal chiller. At higher electrical cost the half-effect cycle has the advantage because of
130

lower electrical operating costs. Figure 8-13 shows that the half-effect cycle requires a cheap
source of fuel to overcome the higher capital cost and larger cooling tower.

Figure 8-13. Comparison between half-effect cycle and electric centrifugal chiller.

At 200
o
F and 0.1 $/kWh the break-even cost of heat is approximately 0.13 $/therm.

The right axis on Figure 8-13 depicts if the heat source were free, what the heat recovery
equipment would have to cost in order to provide the energy [refer to equation 6.5 in section
6.4]. For example, at 0.1 $/kWh and 200
o
F, the heat recovery equipment must cost less than
$1.2 million in order for the project to be feasible [20 year life cycle 4380 hours of operation
per year]. The term heat recovery equipment encompasses all the equipment needed to
transfer the waste heat into a form that can be used by the absorption chiller.
131

The break-even cost of heat is constant for 180
o
F hot water temperatures and higher. The
reason for this is depicted in Figure 8-14, which shows the COP and capacity as a function of
heat source temperature.

Figure 8-14. Change in capacity and COP as a function of entering hot water firing
temperature.

Even though capacity is decreasing, the COP remains constant because the amount of heat
input is also decreasing.

132

8.6 Variation of Capacity due to different operating conditions and designs.

The half-effect was designed for 200
o
F at 2000 gpm, but how does the performance change if
the temperature of hot water flow rate is varied? Figure 8-15 displays a graph of the capacity
due to different hot water flow rates. The change in the capacity is gradual due to a decrease
in hot water flow rate. As the temperature of the hot water decreases so does the capacity.
Therefore, an increase in UA for the generator or absorber is needed to achieve 600 tons for
hot water temperatures less than
o
F.

Figure 8-15. Capacity effect due to different hot water flow rates.

133

The right axis in Figure 8-15 represents the temperature drop across both generators. For
example, at 200
o
F and 2000 gpm the temperature leaving the LT generator would be
approximately 180
o
F, which corresponds to a 20
o
F drop across both generators.

The cooling water flow rate first enters the HC absorber, then the LC absorber and finally the
condenser. In Figure 8-16 the first number represents the flow rate through the absorbers and
the second number is the flow rate through the condenser. Figure 8-16 shows the effect of
changing the flow arrangement by splitting the flow into two streams, one stream enters the
absorber combination and the goes directly to the condenser.

Figure 8-16. Comparison of different cooling water flow arrangements.

134

The base case is 85
o
F at 3100 gpm and the two split cases are 2100 and 2600 gpm into the
absorber and 1000 and 500 gpm into the condenser. According to Figure 8-16 the best
design is the base case in which the water flows in series through all three components.

The internal solution flow rate leaving the absorber can be adjusted using a variable speed
pump to control capacity at part load conditions. The reason for this adjustment instead of
decreasing the hot water flow rate or temperature is because the absorption chiller may be
connected in a co-generation system where the heat source has a constant waste heat flow
rate and requires a certain temperature on the return. Figure 8-17 and Figure 8-18 displays
the effect on capacity for decreasing and increasing the flow rate leaving the absorbers.

Figure 8-17. Effect on Capacity due to varying HC absorber solution flow rate.

135

Figure 8-18. Effect on Capacity due to varying HC absorber solution flow rate.

The discontinuity in the slope in Figure 8-17 and Figure 8-18 is due to the outside heat
transfer is being set to a specific value at and above a certain film Reynolds number [refer to
section 3.4.3] . The left of the knee represents a heat transfer coefficient that is decreasing
with film Reynolds number and the right is a constant heat transfer coefficient with an
increase in film Reynolds number. Figure 8-17 and Figure 8-18 demonstrates the effect on
capacity for independent changes in solution flow rate. Figure 8-19 shows the effect of
reducing both internal solution flow rates by the same percentage.

136

Figure 8-19. Simultaneous reduction in both internal solution flow rate.

The x-axis of Figure 8-19 is the fraction that is multiplied by the design solution flow rate.
For example at 0.7 the HC solution flow rate is 168 gpm instead of 240 gpm. Reducing both
flow rates did not produce a large change in the reduction of capacity compared to Figure
8-17 and Figure 8-18. For example, in Figure 8-17 at 170 gpm, the capacity is approximately
525 tons. Whereas in Figure 8-19 at 0.7 fraction, the capacity is approximately 500 tons. The
only advantage of reducing both flow rates is that the HC solution flow rate could be reduced
more than if the LC was held constant. An important parameter to keep track of when
reducing flow rate is the temperature margin. The right axis displays the change in
temperature margin as both flow rates are reduced. Operating the absorption unit with a
small [5-10
o
F] temperature margin is not an appropriate design because of potential
problems with crystallization.
137

Even though the capacity is reduced in Figure 8-17 and Figure 8-18, it may not be enough for
part load conditions. Another option is to increase the temperature of the cooling water to
the absorber. This will have two effects: it will reduce capacity for part load and reduce fan
power at the cooling tower. The different lines in Figure 8-20 represent different cooling
water flow rates. By increasing the cooling water temperature, capacity can be reduced to
about 450 tons. Another advantage of this technique is that the temperature margin for this
analysis did not fall below 30
o
F.

Figure 8-20. Effect on capacity due to varying cooling water temperature.

138

8.7 Summary

The following is a list of major conclusions regarding the analysis of the half-effect cycle in
Chapter 8.

The same optimum solution of tube fraction in the generator and absorber exists
regardless of series or parallel flow arrangement with 200
o
F water at 2000 gpm. If the
hot water temperature is lowered, the f
abs
changes from 0.41 to a number greater than 0.5,
which means that the LC absorber is now larger than the HC absorber. On the contrary,
the f
gen
did not change.

Similar to the findings for the single-effect cycle as the hot water flow rate increases less
capital investment is required.

The higher the hot water flow rate the lower the entering hot water temperature can go
before there is a difference between series and parallel flow arrangement. At
approximately 190
o
F it is more economical to use series flow rather than parallel flow in
the generator design.

The addition of the cooling tower doubles the cost of the absorption chiller system.

The half-effect cycle requires a cheap source of fuel to overcome the high capital cost
and larger cooling tower. At 200
o
F and 0.1 $/kWh, the break-even cost of heat is
approximately 0.13 $/therm.

The best design for cooling water flow arrangement is to have the water flow in series
through the HC, followed by the LC, and then finally the condenser.

Reducing both internal solution flow rates out of the HC and LC absorber did not
produce a large change in the reduction of capacity compared to the reduction in the
individual flow rates. The capacity can be reduced by 150 tons before there are problems
with the temperature margin becoming to low.

Another option to reduce capacity is to increase the temperature and/or decrease the
solution flow rate of the cooling water. This will have two positive effects: it will reduce
capacity for part load conditions and reduce fan power at the cooling tower, which lowers
operating cost.
139

140

CHAPTER 9

COMPARISON BETWEEN HALF-EFFECT AND SINGLE-EFFECT
ABSORPTION CHILLERS

9.1 Overview

The purpose of Chapters 7 and 8 were to examine the single and half-effect cycle. Chapter 9
compares the two cycles based on economics and capacity. The process is based on finding
what operating conditions are favorable in the single-effect cycle compared to the half-effect
cycle.

9.2 Dollar per ton for different heat source temperatures.

Figure 9-1 and Figure 9-2 compare the capital cost between the single and half-effect
absorption chiller unit and system cost. The system cost consists of the absorption chiller
and cooling tower. The analysis was based on holding capacity constant and allowing the
generator size to vary with entering hot water temperature. Both figures demonstrate that the
single-effect unit has a lower price per ton at firing temperatures above 200
o
F.
141

Figure 9-1. Capital dollars per ton cost comparison between single and half-effect chiller.

Figure 9-2. System dollars per ton cost comparison between single and half-effect chiller.
142

The capital cost is based on material, labor, and overhead. At around 220
o
F, Figure 9-1 has a
cost difference of about 90 $/ton for 2 times the fixed price and 30 $/ton for the equal fixed
price. The system cost difference in Figure 9-2 is 180 $/ton for 2 times the fixed price and
120 $/ton for equal fixed price. An important result of Figure 9-1 and Figure 9-2 is that in
the range of 195-200
o
F at 2000 gpm, there is a transition as to which type of unit one would
want to install. The assumption of fixed price does not change the transition temperature
because of the steep slope for the SE unit. The lower limit based on economics for the half-
effect cycle is about180
o
F at 2000 gpm and 200
o
F for the single-effect unit. Table 9-1
compares the capital cost between the single, half, and electric centrifugal chiller.

Table 9-1. Capital cost comparison between single-effect, half-effect, and electric centrifugal
chiller [600 tons, 200
o
F, 2000 gpm heat source]
Chiller $/Ton Chiller $/Ton
System
Cost range of
Cooling
Tower
Fraction of
Cooling
Tower Cost
Single-Effect
[Base Case]
170 363 55 0.53
Single-Effect
[finned]
185 374 54 0.51
Single-Effect
[smooth]
222 411 54 0.46
Half-Effect 275 550 78 0.50
Electric Centrifugal 150 285 39 0.47

The cooling tower cost for the electric centrifugal chiller is approximately 47% of the total
system cost whereas the cooling tower is about 50% of the total system cost for the
absorption unit. Even for the electric centrifugal chiller the cooling tower is a major cost for
143

the system but the capital cost is lower. The cost range for the cooling tower is used to
compare a galvanized or stainless steel constructed cooling tower.

Figure 9-3 and Figure 9-4 compares the capital and system cost for various hot water flow
rates. The main conclusion from both graphs is that as the hot water flow rate is decreased,
the transition to the half-effect cycle occurs at higher hot water firing temperatures. For
example, at 2000 gpm the transition temperature to the half-effect is 195
o
F, but at 1000 gpm
the temperature is 205
o
F.

Figure 9-3. Capital dollars per ton for different hot water flow rates.

144

Figure 9-4. System dollars per ton for different hot water flow rates.

Figure 9-5 compares the capacity between the SE and HE cycle designs for various hot water
flow rates and temperatures. Both the HE and SE at 200
o
F and 2000 gpm have a capacity of
600 tons. As the heat source flow rate is lowered, the temperature at which the transition to
the HE having a higher capacity occurs at higher temperatures. This means the HE can
achieve a higher capacity than the SE over a larger range of low-temperature and low flow
rate applications.

145

Figure 9-5. Capacity comparison between SE and HE cycle.

In other words, the HE cycle can maintain a higher capacity compared to the SE unit at lower
operating temperatures and flow rates.

9.3 Break-even cost comparison.

Figure 9-6 compares the capacity and COP for the single and half-effect unit. At 200
o
F there
is a transition to the half-effect unit having a higher capacity, which corresponds to the
results in Figure 9-1. The COP for the half-effect unit is linear in the temperature range of
160-220
o
F, whereas at 185
o
F the COP starts to decrease for the SE unit.
146

Figure 9-6. Comparison of capacity and COP for the single and half-effect absorption unit.

Another important result of Figure 9-6 is that the COP for the half-effect unit is about of
the single-effect. This result was seen in Figure 9-2, due to the larger cooling tower and is
also present in Figure 9-7.

147

Figure 9-7. Comparison of cost of heat for the single and half-effect absorption unit.

The break-even cost of heat for the half-effect cycle is about of the single-effect for the
same entering hot water temperature. For example, at 200
o
F and 0.1$/kWh, if heat can be
supplied at approximately 0.13 $/Therm then half-effect cycle has an economic advantage
over the electric centrifugal chiller, whereas it is approximately 0.28 $/Therm for the single-
effect cycle.

148

Table 9-2. Compares the life-cycle cost between the single and half-effect absorption unit.
[20 year life cycle , 600 Tons, 0.05$/kWh, 200
o
F Hot Water @ 2000 gpm, 4380 Hours].
SE Absorption/ Tower
0.14 $/Therm
HE Absorption/ Tower
0.05 $/Therm
Capital Cost
Chiller $136,048 $206,526
Tower $170,591 $248,654
$306,639 $455,180
Operating Cost
Pumps $11,430 $19,600
Heat Source $642,100 $447,360
Tower $72,150 $115,360
$725,680 $582,320
Total $1,032,319 $1,037,500

The only reason the total cost is nearly the same is that the break-even cost of heat is
different. Table 9-2 demonstrates the higher capital cost required for the half-effect cycle:
the main disadvantage of the half-effect cycle is that it is approximately 1.5 times more than
the single-effect, the cooling tower for the half-effect is also about 1.5 times more than the
single-effect, and the half-effect has a higher operating cost if the heat source were supplied
at the same price. The only situation in which the half-effect cycle has an advantage is if the
heat source is free and it is lower than 200
o
F or has a low flow rate.

Table 9-3 examines the difference in UA and tube count for each component of the single
and half effect design. The main difference in the design is the number of generator tubes.
The half-effect cycle is able to produce the same cooling effect with two smaller generators
and a slight increase in the absorber. The total area is less for the half-effect cycle, however t
the capital cost is still greater [Figure 9-1] because of the extra solution pump, heat
exchanger, brine solution, increase in controls due to extra components, etc.
149

An interesting aspect is that the half effect cycle is able to generate the same cooling effect
with approximately twice as much heat input with smaller generators to transfer the heat.
The smaller generators are interesting, but twice as much heat is still needed. Even if it is free
from co-generation, it still puts a burden on the system. Table 9-3 also shows that twice as
much heat must be rejected to the absorber, which explains the need for a larger cooling
tower.
Table 9-3. Tube and UA comparison between half and single effect unit. [600 Tons, 200
o
F
hot water at 2000 gpm.]
Tubes UA [Btu/hr-F] Energy
[Mbtu/hr]
Internal
solution flow
rate[gpm]
Absorber SE 554 558751 9.9 230
Absorber HC, HE 350 375285 9.3 240
Absorber LC, HE 243 327870 8.9 210

Generator SE 1133 / 300 993439 10.7
Generator HE, LT 255 283976 9.7
Generator HE, HT 192 246966 8.9

Condenser SE 128 882375 8.0
Condenser HE 128 852738 7.6

Evaporator SE 336 1.354e6 7.2
Evaporator HE 336 1.35e6 7.2

The ability to have smaller generators transferring the same amount of heat compared to the
SE generator is due to the LT generator operating at a mid-pressure and the HT generator
operating at a lower concentration. In summary, a mid pressure with a smaller surface area is
as effective as a high pressure with a larger surface area.

150

9.4 Summary

The following is a summary of the comparison between the single and half effect absorption
chiller system

The single-effect unit has a lower price per ton at firing temperatures above 200
o
F.

In the range of 195-200
o
F at 2000 gpm there is a transition from the single-effect unit to
the half-effect unit being installed.

For all systems including the electric centrifugal chiller, the cooling tower is about 50%
of the total system cost. The disadvantage for absorption unit is that the capital cost is
higher because of the lower COP.

The half-effect cycle can achieve a higher capacity compared to the single-effect over a
larger range of low temperature and low flow rate applications.

The only way in which the half-effect cycle is advantageous compared to over the single-
effect cycle is if the heat source is relatively free and it is lower than 200
o
F or has a low
flow rate.

The break-even cost of heat for the half-effect cycle is about the cost of the single-
effect for the same entering hot water temperature.

Desorption is most effective at lower pressures and lower concentrations, which explains
why the half-effect cycle can operate with smaller generators compared to the single-
effect unit.
151

152

CHAPTER 10

CONCLUSIONS AND RECOMMENDATIONS

10.1 Conclusions

The following is a list of important conclusions from the study of low-temperature absorption
chiller systems.

From examining the published literature, the single-effect unit finds applications in co-
generation, geothermal energy, and combined heating and cooling plants. The best
method for designing a single-effect unit to operate on low heat source temperatures is to
examine the heat exchanger area of all components.

Holding capacity constant [600 tons], at 205
o
F the base case chiller size starts to change
rapidly in generator heat exchanger area to accommodate the lower firing temperatures.

Based on the assumption that finned tubes cost twice as much as smooth tubes, the
reduction in generator tubes using finned tubes is enough to compensate for the higher
price.

The new low temperature single-effect absorption unit was designed by increasing the
number of tubes in the generator [262 300 ], using finned tubes instead of smooth
tubes, increasing the number of tubes in the absorber [514 554 ], and increasing low
temperature heat exchanger effectiveness [0.76 0.8.].

The new low temperature single-effect hot water design has an increase of 15 $/ton to
obtain a capacity of 600 tons with 200
o
F, compared to 227
o
F hot water at 2000 gpm.

The optimum value for the cooling water flow rate is independent of the heat source
temperature.
153

The cost of the absorption chiller system for the single and half-effect has approximately
doubled due to the cooling tower. This is also true for an electric centrifugal chiller but
the capital investment for the cooling tower in the SE is $170,000 compared to $80,000
for the electric chiller.

If the heat source can be supplied at a cost less than or equal to approximately 0.28
$/Therm when electricity is 0.1 $/kWh then absorption is the preferred alternative [based
on 600 tons capacity with 2000 gpm hot water].

The half-effect cycle requires a cheap source of fuel to overcome the high capital cost
and larger cooling tower. At 200
o
F and 0.1 $/kWh, the break-even cost of heat is
approximately 0.05 $/therm.

The cost of heat for the absorption chiller is 90% of the total operating cost over one year.
Thus, using absorption with co-generation is one of the only feasible situations where it
would be more economically competitive than current centrifugal chillers. The other
situation is high electrical cost.

The best design for the cooling water flow arrangement in the half-effect cycle is to have
the water flow in series through the HC, followed by the LC, and finally then the
condenser.

One option to reduce capacity in the half-effect cycle is to increase the temperature
and/or decrease the solution flow rate of the cooling water. Using cooling water at 95
o
F,
2100 gpm reduces the capacity to 450 tons. This technique will have two positive
effects: it will reduce capacity for part load conditions and reduce fan power at the
cooling tower, which lowers operating cost. Another option is to use variable speed
pumps for the HC and LC absorber solution flow rates. Decreasing both solution flow
rates can lower the capacity to 450 tons without the temperature margin decreasing below
10
o
F.

The single-effect unit has a lower price per ton compared to the half-effect cycle for
firing temperatures above 200
o
F.

In the range of 195-200
o
F at 2000 gpm there is a transition from the single-effect unit to
the half-effect unit being installed.

The half-effect cycle can achieve a higher capacity compared to the single-effect cycle
over a larger range of low temperature and low flow rate applications.

The only advantage of the half-effect cycle over the single-effect cycle occurs if the heat
source is free and it is lower than 200
o
F or has a low flow rate.

154

The break-even cost of heat for the half-effect cycle is about the cost for the single-
effect for the same entering hot water temperature.

The total number of tubes for the half-effect cycle is less then the single-effect cycle even
though the heat transfer in the generator and absorber is twice as much. This result is due
to the fact that desorption is most effective at lower pressures and lower concentrations.
The half-effect unit cost more because it requires at minimum one additional pump,
another brass plate heat exchanger, additional brine solution, additional control
equipment, pipingetc.

10.2 Recommendations

Each component was modeled based on an inside and outside heat transfer coefficient. One
detail of further study would be to examine the effect of changing the tube arrangement in
each component. This analysis would involve determining the dependence of tube
arrangement on the outside heat transfer coefficient. For example, in the condenser, as the
fluid condenses on the pipes, there is an increase in the thickness of the liquid as it falls down
the pipe. This has the effect of decreasing the heat transfer coefficient along a tube column.

With the completion of designing the single and half-effect cycle for low temperature
applications, an analysis with case studies would give another perspective on possible
applications. The analysis of this paper involved looking at constant cooling loads for 8760
hours per year and 4380 hours per year. A case study involving a building with a cooling
and electrical load would help determine the viability of an absorption unit for co-generation.
Another approach would be to contact industries to determine their waste heat output and if
there is a need for cooling in the area.
155

156

REFERENCES

Electrical Research Association, (1967). 1967 Steam Tables, Thermodynamic Properties of
Water and Steam; Viscosity of Water and Steam, Thermal Conductivity of Water and Steam,
Edward Arnold Publishers, London.

Beckman W. A., and Duffie J. A. (1991). Solar Engineering of Thermal Processes, John
Wiley & Sons.

Braun J. E., Klein S. A., and Mitchell J. W. (1989). Effectiveness models for cooling towers
and cooling coils. ASHRAE Trans., 96, 164-174.

Bruno J.C., Fernandez F., and Castells F. (1996). Absorption Chillers Integration in a
Combined Heat and Power. Ab-Sorption, 2, 759-767.

Marley Cooling Tower Company, (1998). UPDATE [Unitary Product Data and Thermal
Evaluation]. , Overland Park.

Cosenza F., and Vliet G.C. Ph.D. (1990). Absorption in Falling Water/LiBr Films on
Horizontal Tubes. ASHRAE Transaction, 96, 693-701.

Erickson D.C., (1995). Waste-Heat-Powered Icemaker for Isolated Fishing Village.
ASHRAE Transactions, 101(Part 1), 1185-1188.

Fairchild M., (2000). Marley Cooling Tower Company.

ASHRAE Handbook, (1989). Fundamentals I-P. , American Society of Heating,
Refrigerating, and Air-Conditioning Engineers, Inc., Atlanta.

Herold K.E., Radermacher R., and Klein S.A. (1996). Absorption Chillers and Heat Pumps,
CRC Press.

Homma R., Nishiyama N., and Wakimizu H. Simulation and Experimental Research of
Single-Effect Absorption Refrigerators Driven by Waste Hot Water. Proceedings of the
International Absorption Heat Pump Conference, New Orleans, Louisiana, 273-278.

IEA District Heating and Cooling (1999). District Cooling, Balancing the Production and
Demand in CHP. Novem, Netherlands, 1-180.

RS Means Co. Inc., (1998). Mechanical Cost Data HVAC & Controls, .

Incropera F.P., and DeWitt D.P. (1996). Fundamentals of Heat and Mass Transfer, John
Wiley & Sons.
157

Kececiler A., Acar H., and Dogan A. (1999). Thermodynamic analysis of the absorption
refrigeration system with geothermal energy: an experimental study. Energy Conversion
and Management, 41(1), 37-48.

Koeppel, E. A. (1994). The Modeling, Performance, and Optimal Control of Commercial
Absorption Chillers, MS, UW Madison, Madison.

Lamp P., and Ziegler F. (1996). Comparison of different liquid and solid sorption systems
with respect to low temperature driving heat. Ab-Sorption, 1, 269-276.

MA W.B., Xia H.W., and Deng S.M. Industrial Application of a Two-Stage, Half Effect
LiBr/H20 Absorption Chiller. Ab-Sorption, Montreal, Quebec, Canada, 679-683.

U.S Chiller Manufacturer, (2000).

Meloche N., Larsson I., and Snoek C. (1996). Integrating District Cooling with Combined
Heat and Power. Ab-Sorption, 2, 795-802.

Plzak, B. Absoprtion Cooling Demonstration and Application. Ab-Sorption, Montreal,
Quebec, Canada, 67-74.

Rohensenow W.M., Hartnett J.P., and Cho Y.I. (1998). Handbook of Heat Transfer,
McGraw-Hill Handbooks.

Schweigler C., Demmel S., and Ziegler F. Single-Effect/Double-Lift Chiller: Operational
Experience and Prospect. Proceedings of the International Sorption Heat Pump Conference,
Munich, Germany, 533-539.

Thornbloom M.D., and Nimmo D.G. Modification of the Absorption Cycle for Low
Generator Firing Temperatures. Solar Engineering, San Francisco, CA, 367-372.

Wang C., Zhen Lu, and Zhou J. Enhancement of Heat and Mass Transfer in Lithium
Bromide Falling Film Generator. Proceedings of the International Sorption Heat Pump
Conference, Munich, Germany, 301-305.

158

APPENDIX A CD-ROM FILES

EES Code
The folder EES code contains the model for the single and half-effect absorption chiller. The
folder also contains the component subroutines that the single and half-effect model calls.
Functions
Tseq.txt Equilibrium temperature of LiBr-water solution
Tveq.txt Equilibrium temperature of Water
reynolds_dg.txt Reynolds number
libr_cp.txt Specific heat for LiBr-water solution
libr_con.txt Conductivity for LiBr-water solution
libr_den.txt Density for LiBr-water solution
libr_vis.txt Viscosity for LiBr-water solution
xeq.txt Equilibrium concentration for LiBr-water solution
libr_Tc2.txt Crystallization temperature using Trane UCP2 controller
libr_Tcry.txt Crystallization temperature using Foote Mineral Co.
correlation
hfo_abs.txt Outside heat transfer coefficient for absorber
UALTHX.txt Brass plate heat exchanger model
massflow.txt Internal mass flow rate leaving the absorber control
generatorNTU.txt Calculates NTU for cross or counter flow heat exchanger
CoolingTowerSelection.txt Different cooling tower parameters.
Generatortype.txt Control for generator type, steam or hot water.
tubetype.txt Control for generator tube type, smooth or finned tube

Modules

CoolingTowerMod.txt Cooling tower model for absorption unit
CoolingTowerModVC.txt Cooling tower model for electric chiller
generator.txt Absorption generator UA model
absorber.txt Absorption absorber UA model
condenser.txt Absorption condenser UA model
evaporator.txt Absorption evaporator UA model
vaporcompression.txt Vapor compression model, not used but still good
pump.txt Pump model used to calculate the pump power

159

Cooling Tower Performance Curves

In the Marley folder there is a program named Update.exe that provide performance
curves for a specific cooling tower. The program can also be used to determine the fan
power required for a specified capacity.

Electronic Word Document

An electronic copy of the masters thesis is provided.

Miscellaneous Folder

The folder contains EES, Excel and Word programs. For example an EES program to
produce Dhring plots.

160

Goodheart KA

Enviado por

Dados do documento

Descrição original:

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Goodheart KA

Enviado por

Direitos autorais:

Formatos disponíveis

Low Firing Temperature Absorption Chiller System

Você também pode gostar